Library · Signals, sensing & imaging
What Image Segmentation & Computer Vision got wrong, from 99 theses
Summary written by a model from the records below: The records document numerous empirical failures across computer vision pipelines, ranging from thresholding heuristics to deep neural network segmentation architectures. Common difficulties include vulnerability to domain shift, extreme latency during dense inference, boundary confusion among overlapping instances, and underperformance compared to simpler baselines. These records come from graduate theses, 2021 to 2026, each labeled with its degree where the repository states it, and each links to its thesis. They were extracted by language models reading the full text, so treat each as a lead to read, not a verdict.
Plus 10 problems the authors left open, listed at the end.
Failure modes, by theses reporting them
Select a mode to jump to its records.
Complex segmentation architectures and automated pipelines fall short of simpler baselines and manual contouring
Advanced deep learning models and multi-stage pipelines frequently achieved lower overlap scores, lower recall, or worse boundary stability than standard baselines and manual delineations. Automated segmentation systems also struggled with reliability on difficult cases, requiring manual contour corrections to prevent diagnostic or procedural errors.
Lost to a baseline
Single-class semantic segmentation achieved 85% Mean IoU, whereas two-class training achieved lower Mean IoU (80%) due to irregular, jagged predicted boundaries on metric bars.
Integration and classification of spatial data for 3D modelling and monitoring of built heritage · Politecnico di Torino · PhD
Lost to a baseline
A simple 1% labeled-frame threshold baseline beat the complex combined labeling-and-segmentation dual-threshold method in prognostic stratification (85.96% vs 83.33%).
Development of Lung Ultrasound Quantitative Approaches and Automatic Semi-Quantitative Strategies: In Silico, In Vitro, and Clinical Studies · University of Trento · PhD
Lost to a baseline
Under outlier perturbations, baseline Binary Segmentation (best-case and default) outperformed Repeated Binary Segmentation in stability.
Stability of change point detection methods · Imperial · Master's
Lost to a baseline
Original SAM achieved higher recall (0.53 vs 0.41) than the fine-tuned cropping approach on bare cropland, though driven by over-segmentation/false positives.
A Unified Framework for Advancing Soil Erosion and Flood Assessment Through Deep Learning and Process-Based Modeling · Tübingen · PhD
Lost to a baseline
Direct 2D segmentation without priors (Deep-Single) had higher surface error outliers on noisy B-scans compared to the graph-based Aura baseline.
RETINAL OCT IMAGE ANALYSIS USING DEEP LEARNING · Johns Hopkins · PhD
Lost to a baseline
ResNet with ASPP module on radargram segmentation achieved overall accuracy slightly lower than literature SVM baselines
Advanced methods for simulation-based performance assessment and analysis of radar sounder data · University of Trento · PhD
Lost to a baseline
Epistemic uncertainty data selection (Hausdorff 4.15 mm) lost to baseline fully supervised (Hausdorff 3.97 mm) in MRI segmentation
Improving deep-learning segmentation performance in 3D neuroimaging with minimal manual annotations · Oxford · PhD
Lost to a baseline
Adv-trained ResDSN Coarse 3D medical image segmentation performance on clean data (79.09% accuracy/Dice) lost to baseline ResDSN Coarse (87.84%).
TOWARDS DEEP LEARNING ROBUSTNESS FOR COMPUTER VISION IN THE REAL WORLD · Johns Hopkins · PhD
Lost to a baseline
DeepLabv3 achieved lower accuracy metrics (F1 score 0.751 vs 0.809) on visible image segmentation compared to FPN despite having over 2x the parameter count.
Condition Assessment of Civil Infrastructure and Materials Using Deep Learning · Virginia Tech · Doctorate
Lost to a baseline
SegResNet demonstrated slightly higher sensitivity in overall liver tumor segmentation than SmoothSegNet despite having lower accuracy and Dice score due to over-segmentation.
Knowledge-Informed Weakly-Supervised Deep Learning Models for Cancer Applications · Georgia Tech · PhD
Lost to a baseline
Def-RgDL SMG segmentation (DSC 53.8%) was inferior to DL alone without guidance (DSC 63.5%) in a case where registration propagated a false vessel contour.
Efficient and Intelligent Radiotherapy Planning and Adaptation · UT Southwestern · PhD
Lost to a baseline
SDXL-Turbo achieved lower semantic segmentation performance (36.99% overall Dice-AUC) compared to full SDXL (38.27% overall Dice-AUC).
METHODOLOGICAL ADVANCES IN REAL-WORLD COMPUTER VISION: ROBUSTNESS, EFFICIENCY AND HUMAN PREFERENCE ALIGNMENT · Cornell · PhD
Considered and rejected
Considered and rejected: Rejected using a dual 3D segmentation head ensemble for domain generalization in LiDOG as it performed worse than the auxiliary 2D BEV projection decoder (30.07 vs 44.18 mIoU).
Handling Domain Shift in 3D Point Cloud Perception · University of Trento · PhD
Lost to a baseline
U-Net beat U-KAN on Sentinel-1 crop field segmentation Recall (85.56% vs 77.50%)
Spatio-Temporal Machine Learning for Ecology and Crisis Management · Politecnico di Torino · PhD
Lost to a baseline
Manual clover dry matter fraction prediction from Hansen et al. [101] achieved 7.8% standard deviation, surpassing the thesis's ERFNet grass/legumes segmentation in complex outdoor lighting
Tried and failed
multimodal OCR preprocessing for vision-language models applied to code screenshot understanding. Outcome: worse than baseline. Reason: explicit OCR features provided no measurable improvement over native visual understanding in large multimodal models
Use of Visual Content for Inference and Response in Q/A Forums · York University · Master's
Lost to a baseline
Manual full-organ segmentation outperformed the simpler, faster ROI-based baseline method in test-retest repeatability across nearly all sequences.
Lost to a baseline
ViNet consistently outperformed DReyeVENet (which utilized optical flow and semantic segmentation) and BDD-ANet (which used weighted sampling) on DR(eye)VE and BDD-A benchmarks.
Investigating and Modeling the Effects of Task and Context on Drivers' Attention · York University · PhD
Considered and rejected
Considered and rejected: Applying test-time OOD detection algorithms (other than simple MSP) to Vision Transformers (ViTs) due to poor performance compared to ConvNets
A Study of Techniques for Robustness to Out-of-Distribution Examples · Dalhousie · PhD
Considered and rejected
Considered and rejected: Rejected ResNet-50 backbone for final hyperbolic metric learning publication because vision transformers consistently outperformed CNNs.
Self-supervised Representation Learning in Computer Vision and Reinforcement Learning · University of Trento · PhD
Considered and rejected
Considered and rejected: Excluded automated computer-vision machine learning classification in favor of manual human coding due to current ML limitations in recognizing complex semantic scene context.
A Collective Sense of Place and the Image of the City @ Urban Public Spaces: Analysis on People's Perception of User-Generated Image Content and Hashtags on Instagram · Virginia Tech · Doctorate
Lost to a baseline
Automated nnU-NET SAT segmentation suffered complete or partial failure in 13% of datasets, requiring manual correction compared to reliable manual masking.
Magnetic Resonance Imaging and Spectroscopy Methods for Studying Obesity: Applications for Bariatric Surgery · Nottingham
Lost to a baseline
CardioINSIGHT automated CT segmentation inconsistently failed compared to manual multi-slice 2D contour interpolation.
Feasibility of improving risk stratification in the inherited cardiac conditions · Imperial · PhD
Lost to a baseline
CoBEVT achieved 60.4% mIoU on vehicle BEV segmentation and 63.0% on drivable area, beating CoBEVFusion (59.5% and 61.7% mIoU)
Enhancing Perception for Autonomous Vehicles · Queen's University · PhD
Lost to a baseline
K-means achieved higher Precision in traffic frame segmentation (0.89 vs 0.87 for HHGATSD)
Novel optimization methods and model for improving sustainability and efficiency in last-mile logistics · University of Deusto · PhD
Models fail to generalize across domain shifts, differing resolutions, and synthetic datasets
Direct transfer across imaging protocols, slice resolutions, and synthetic-to-real environments caused severe under-segmentation, discontinuous predictions, and high false positive rates. Techniques like mixup augmentation, adversarial bias field simulation, and synthetic data generation frequently degraded performance on real out-of-distribution targets.
Tried and failed
U-Net segmentation cross-dataset transfer applied to histology image segmentation. Outcome: did not generalise. Reason: Domain shift from differing staining and imaging protocols reduced accuracy to simple thresholding baseline
Computational modelling of diffusion magnetic resonance imaging based on cardiac histology · Imperial · PhD
Tried and failed
direct cross-resolution neural network deployment applied to medical image segmentation. Outcome: did not generalise. Reason: resolution and signal-to-noise mismatch between high-resolution multi-average training data and lower-resolution single acquisitions
Deep learning-based analysis of multiple sclerosis lesions with high and ultra-high field MRI · EPFL · PhD
Tried and failed
2D transfer learning segmentation on 3D volumes applied to low-resolution medical image segmentation. Outcome: did not generalise. Reason: Reduced signal and resolution caused severe sensitivity loss for small target regions below a size threshold.
Machine Learning And Quantitative Neuroimaging In Epilepsy And Low Field Mri · Penn · PhD
Tried and failed
concatenating multi-timestep diffusion features for classification applied to semi-supervised image segmentation. Outcome: did not generalise. Reason: concatenation overfits in low-data regimes, degrading out-of-domain performance compared to ensembled predictions
Label-efficient medical image segmentation: the benefits and limitations of semi-supervised generative models · Imperial · Doctorate
Tried and failed
generative neural network for data augmentation applied to imbalanced 3D medical image segmentation. Outcome: did not generalise. Reason: generated augmentations caused heavy overfitting to validation data and failed to generalize to unseen test data
Learning strategies for improving neural networks for image segmentation under class imbalance · Imperial · PhD
Tried and failed
off-the-shelf object segmentation without domain-specific fine-tuning applied to surgical video anatomy detection. Outcome: did not generalise. Reason: models pre-trained on common natural images fail to recognize specialized surgical anatomy
XMARCUS: A Pathway Towards Remote Robotic Surgery Coaching · Virginia Tech · Master's
Tried and failed
deep learning auto-segmentation for automated quality assurance applied to clinical target volume delineation. Outcome: did not generalise. Reason: Protocol variability, error propagation from upstream inputs, and inability to encode subjective clinical judgment.
INTEGRATION OF ARTIFICIAL INTELLIGENCE IN QUALITY ASSURANCE FOR HEAD AND NECK CANCER RADIOTHERAPY CLINICAL TRIALS · Penn · PhD
Tried and failed
direct transfer of synthetic-trained 3D CNNs applied to real 3D volumetric image segmentation. Outcome: did not generalise. Reason: domain gap between synthetic training data and noisy real data caused discontinuous predictions and low recall
Multiscale Integration of Cross-Modal Subsurface Data for Reservoir Characterization under Label-Constrained Environments · Georgia Tech · PhD
Tried and failed
direct cross-dataset model transfer without domain adaptation applied to electron microscopy 3D segmentation. Outcome: did not generalise. Reason: acquisition variations across datasets caused severe undersegmentation and high merge errors
Grounded: Inference via Local Signals & Learned Representations by Organic & Artificial Systems · Harvard · PhD
Tried and failed
handcrafted feature clustering for image segmentation applied to medical lesion segmentation. Outcome: did not generalise. Reason: moderate segmentation overlap and vulnerability to heavy illumination shadows
Tried and failed
single-modality deep learning auto-segmentation for error screening applied to radiotherapy target volume contour validation. Outcome: did not generalise. Reason: CT-only segmentation produced an unacceptably high false positive rate for anomaly detection
INTEGRATION OF ARTIFICIAL INTELLIGENCE IN QUALITY ASSURANCE FOR HEAD AND NECK CANCER RADIOTHERAPY CLINICAL TRIALS · Penn · PhD
Tried and failed
mixup data augmentation applied to medical image segmentation across population shifts. Outcome: worse than baseline. Reason: linear interpolation created unrealistic images that degraded generalization across pathological groups
Improving the domain generalization and robustness of neural networks for medical imaging · Imperial · PhD
Tried and failed
adversarial bias field augmentation applied to medical image segmentation domain generalization. Outcome: did not generalise. Reason: increased vulnerability to out-of-distribution spike noise artifacts, degrading segmentation performance
Improving the domain generalization and robustness of neural networks for medical imaging · Imperial · PhD
Lost to a baseline
Heuristic TEA degraded performance for DeepMedic on cross-site prostate segmentation (Site B) compared to no TEA (67.4% vs 71.7% DSC with learned class-specific TRA).
Learning strategies for improving neural networks for image segmentation under class imbalance · Imperial · PhD
Lost to a baseline
Models trained with synthetic long cells showed decreased segmentation performance on in-distribution short cells compared to models trained exclusively on short cells.
Bacterial Deepfakes: Generating Synthetic Microscopy Data to Improve Adaptability of Deep Learning-Based Segmentation Models · University of Washington · PhD
Considered and rejected
Considered and rejected: Training Ultra-TransUNet using ground truth masks as previous frame inputs was rejected in favor of nn-UNet initial segmentation masks to match inference distribution.
Towards Robust Medical Imaging Diagnosis Systems via Integrating Domain Knowledge · Johns Hopkins · PhD
Tried and failed
fine-tuning vision-language models on synthetic text rewrites applied to digital pathology image-text classification. Outcome: worse than baseline. Reason: synthetic rewrites degraded representation quality and context modulation harmed accuracy under non-trivial context selection
Generative AI for Digital Pathology: Advancing Diffusion, Vision-Language, and Safety-Aware Models · Cornell · PhD
Considered and rejected
Considered and rejected: Rejected fully simulated datasets as sole ground truth for defect segmentation because synthetic defects fail to capture real shape complexity, boundary definition, and noise interactions.
End-to-End Artificial Intelligence-Based Pipeline for Quantitative Reporting of Lung V/Q Scintigraphy—VQ-SPRINT: Segmentation, Pseudo-planar Generation, and Registration INtegration Tool · Carleton University · PhD
Considered and rejected
Considered and rejected: Rejected supervised machine-learning organ segmentation trained on external datasets (CT-ORG, AAPM Thoracic) due to domain shifts and severe disease exclusion in prior benchmarks, choosing unsupervised morphological processing instead.
Towards Fully Automated Interpretation of Volumetric Medical Images with Deep Learning · Duke · PhD
Tried and failed
semantic segmentation model trained on urban driving applied to aerial building facade and ground segmentation. Outcome: did not generalise. Reason: domain shift between urban street-level perspective and aerial or campus imagery degraded feature recognition
Aerial Cadastral and Flood Assessment for Disaster Risk Management in Appalachia · Virginia Tech · Doctorate
Tried and failed
synthetic object injection for anomaly segmentation applied to road scene obstacle detection. Outcome: did not generalise. Reason: biases model toward large nearby objects, missing small distant obstacles and causing false positives
Detecting Anomalies and Obstacles in Road Scenes · EPFL · PhD
Semantic segmentation collapses overlapping instances and breaks topological continuity
Pixel-wise semantic segmentation models failed to separate intersecting, crowded, or overlapping structures, frequently merging distinct objects into single masks. Standard pixel-level loss functions and convolutional networks also struggled to preserve continuous thin structures like cracks and fine vessels, yielding fragmented or over-smoothed boundaries.
Tried and failed
centerline and distance-based segmentation loss functions applied to thin curvilinear structure segmentation. Outcome: worse than baseline. Reason: degraded segmentation performance compared to standard Dice loss
Pavement Crack Segmentation with Dense Local Geometry Features and Boundary Enhancement Loss · Georgia Tech · PhD
Tried and failed
fixed margin bounding box cropping and resizing applied to semantic image segmentation. Outcome: worse than baseline. Reason: fixed margin crops either truncate peripheral features or introduce irrelevant distracting background context
Deep face tracking and parsing in the wild · Imperial · PhD
Lost to a baseline
Mesmer whole-cell segmentations produced more rounded boundaries, under-capturing irregular membrane/cytoplasmic extensions compared to traditional watershed segmentation.
Computational Methods to Process and Analyze High-dimensional Imaging Mass Cytometry Datasets in Pathological Tissue Samples · Queen's University · Master's
Lost to a baseline
The proposed hybrid approach yielded lower Recall_Zhang values compared to baseline methods to prioritize under-segmentation prevention
An improved segmentation and classification method for building extraction from RGB images using GEOBIA framework · Queen's University · PhD
Considered and rejected
Considered and rejected: Rejected relying purely on monocular semantic segmentations without instance labels because overlapping obstacles merge into single impassable blocks.
Perception Enabled Planning for Autonomous Systems · Cornell · PhD
Considered and rejected
Considered and rejected: Semantic segmentation via standard U-Net rejected due to inability to differentiate multiple or intersecting instruments without extensive post-processing
Detection and 3D Localization of Surgical Instruments for Image-Guided Surgery · Johns Hopkins · Master's
Considered and rejected
Considered and rejected: Directly using semantic segmentation 'otherprop' class alone for object extraction without plane detection (rejected due to ScanNet under-segmentation causing poor object boundary precision)
Object change detection for autonomous indoor robots in open-world settings · TU Wien · PhD
Considered and rejected
Considered and rejected: Single-stage semantic segmentation network on uncropped images was rejected due to unacceptable false positive rates on internal organelles and out-of-focus boundaries.
Quantifying RTK Signal Transduction Processes Across The Plasma Membrane · Johns Hopkins · PhD
Considered and rejected
Considered and rejected: Diverged from conventional pixel-wise semantic labeling in favor of mask classification via prompt-guided SAM segmentation.
Builderformer : a panoptic segmentation approach to automatic floor plan analysis using vision transformers · University of Moratuwa · Master's
Tried and failed
semantic segmentation of spatially adjacent overlapping features applied to medical image segmentation of multiple findings. Outcome: did not generalise. Reason: models confuse boundary distinctions when multiple distinct localized patterns co-occur in close spatial proximity
AI Systems for Understanding and Grounding Radiology Reports · Harvard · Master's
Lost to a baseline
Custom-trained ilastik pixel classifier underperformed generic pre-trained deep learning models (Cellpose and StarDist) on nuclear segmentation of crowded/touching cells
Exploring tumour heterogeneity and responses to therapy using single-cell resolved microscopy · Imperial · PhD
Considered and rejected
Considered and rejected: Rejected standard CNN local feature extraction alone for overlap segmentation because it loses global context, resulting in subpar segmentation
Cervical Cell Separation using Deep Learning Techniques · Carleton University · Master's
Considered and rejected
Considered and rejected: Rejected frame-wise dense cross-attention (VisTR style) in video segmentation due to slow training convergence and background mask bleeding
Language and Context Guided Object Analysis in Images and Videos · University at Buffalo · PhD
Tried and failed
hierarchical point cloud networks and geometric descriptors applied to fine surface defect segmentation. Outcome: worse than baseline. Reason: they failed to capture subtle fine-grained surface crack details compared to raw RGB and geometric data
3D Segmentation and Damage Analysis from Robotic Scans of Disaster Sites · Georgia Tech · PhD
Tried and failed
hardcoding domain-specific anatomical constraints into neural network applied to image segmentation models. Outcome: worse than baseline. Reason: rigid boundary rules reduced segmentation accuracy and introduced software instability
A study of the relationship between monocotyledonous plant anatomy and water · Nottingham · PhD
Tried and failed
pixel-wise loss functions for continuous line segmentation applied to crack detection in structural images. Reason: standard pixel losses fail to preserve topological continuity, producing fragmented segmentations
Considered and rejected
Considered and rejected: Rejected panoptic segmentation for image region indexing because it is limited to predefined categories and fails to index partial object regions queried by users.
Segmentation-based Image Similarity Search · University of Marburg · PhD
Considered and rejected
Considered and rejected: Rejected stateful panoptic fusion for streaming sectors because it causes under-segmentation for objects spanning into upcoming unseen sectors.
On the Use of Vision and Range Data for Scene Understanding · Johns Hopkins · PhD
High computational complexity and inference latency hinder real-time and 3D segmentation workflows
Applying volumetric three-dimensional neural networks, dense vision transformers, or iterative post-processing steps led to prohibitive computational costs and slow inference speeds. In robotic vision, ultrasound, and video applications, these heavy segmentation architectures were rejected or beaten by faster two-dimensional models, bounding-box detectors, or lightweight baselines.
Tried and failed
real-time one-stage instance segmentation applied to sidewalk semantic segmentation. Reason: produced inaccurate segmentations and exhibited high sensitivity to input image resolution without specific backbone tuning
Robust and precise sidewalk detection with ensemble learning: Enhancing road safety and facilitating curb space management · Iowa State · Master's
Tried and failed
single-stage real-time instance segmentation applied to onboard robotic vision. Outcome: worse than baseline. Reason: produced lower-quality segmentation masks compared to an optimised two-stage instance segmentation model
Autonomous exploration and object reconstruction with an MAV · Imperial · PhD
Tried and failed
pointwise Gaussian process regression applied to point cloud segmentation. Outcome: too slow. Reason: computational cost was too high for real-time processing and struggled with occlusion
Online vehicle trajectory extraction based on LiDAR data · Texas Tech · PhD
Tried and failed
star-convex object detection for tracking and segmentation applied to large-scale high-throughput image sequences. Outcome: too slow. Reason: per-frame inference caused excessive overall processing times on long high-throughput video sequences
Towards systems biophotonics in microscopy, medicine, and robotics · Georgia Tech · PhD
Tried and failed
connected-component analysis post-processing applied to neural network image segmentation masks. Outcome: too slow. Reason: Increased inference time tenfold without providing any meaningful improvement in segmentation accuracy
Automated cardiac segmentation pipeline and motion analysis in rodent models of pulmonary hypertension · Imperial · PhD
Lost to a baseline
MobileNet + U-Net catheter segmentation achieved lower catheter IoU (0.780) than ResNet + U-Net (0.845) on video A training, but was chosen due to inference frame rate (11.04 it/s vs 6.23 it/s)
Deep Learning and Augmented Reality for 3D human-machine interaction · Politecnico di Torino · PhD
Lost to a baseline
Single-image U-Net and DeepLabV3 segmentation models significantly outperformed all fusion architectures (CMNeXt, MMSFormer, StitchFusion) in inference latency (5.0–5.2 ms vs 18.2–83.9 ms).
Autonomous System for Identifying and Capturing Floating Waste · Georgia Tech · PhD
Lost to a baseline
The proposed hybrid segmentation algorithm required double the processing time of LSA and OHRH baseline methods due to dynamic per-loop metric updates
An improved segmentation and classification method for building extraction from RGB images using GEOBIA framework · Queen's University · PhD
Lost to a baseline
VGG-16 semantic segmentation achieved only 2 fps with frequent misidentifications and lost to manual contact determination
Characterisation of a Novel Bioadhesive Used by the Ctenophore Pleurobrachia pileus · University College Dublin · PhD
Lost to a baseline
StarryNite (SN) had faster runtime than the proposed MaskRCNN + optimal transport segmentation framework (though SN achieved lower Recall and only generated centroids).
Computer Vision Methods for Fast and Automated Processing of C. elegans Whole-Brain Videos · Georgia Tech · PhD
Considered and rejected
Considered and rejected: Rejected CNN-based face segmentation models for the real-time wearable implementation due to high computational complexity.
Augmented reality for the sight impaired · Oxford · PhD
Considered and rejected
Considered and rejected: Direct 3D deep learning segmentation rejected due to excessive computational expense and hardware demands for real-time ultrasound imaging, opting for 2D slice-by-slice U-Net.
Forward-Viewing Ultrasound Guidance of a Robotically-Steered Guidewire for Peripheral Interventions · Georgia Tech · PhD
Considered and rejected
Considered and rejected: Rejected instance object segmentation (e.g., Mask R-CNN/UNet) in favor of bounding-box object detection (YOLOv4) due to inference latency and higher data annotation burden.
Vision-Based Force Planning and Voice-Based Human-Machine Interface of an Assistive Robotic Exoskeleton Glove for Brachial Plexus Injuries · Virginia Tech · Doctorate
Considered and rejected
Considered and rejected: Rejected pixel-wise semantic/instance segmentation and GrabCut for object extraction due to high computational overhead in robotics.
Scene understanding via scene graph Szenenverständnis mittels Szenengraph · Leibniz Universität Hannover · PhD
Considered and rejected
Considered and rejected: Rejected 3D CNNs and R-CNN architectures for dynamic frame segmentation because driving video frames lack strict temporal dependence and U-Net was more computationally efficient.
Methods for Classifying Driver Engagement in Autonomous Vehicles Using Physiological Sensors · Carleton University · Master's
Considered and rejected
Considered and rejected: Decided against transformer architectures for semantic segmentation due to extreme computational expense and large data requirements compared to CNNs.
Integration of machine learning for enhanced digital rock physics workflows · UT Austin · PhD
Considered and rejected
Considered and rejected: Decided against semantic segmentation architectures (pixel-wise segmentation) due to high computational cost and because area composition estimation does not require precise item boundary segmentation.
Development of a method to classify and analyse the composition of mixed waste materials in real-time. · Cranfield · PhD
Considered and rejected
Considered and rejected: Rejected standard 3D CNNs from scratch for organ segmentation due to training instability, lack of 3D pre-trained weights, and high computational cost
Towards Robust Deep Learning for Medical Image Analysis · Johns Hopkins · PhD
Unimodal and single-frame visual models produce false positives due to missing context
Relying strictly on individual visual image frames without elevation data, temporal sequences, or multimodal cues led to high false positive rates and spatial dispersion. Models struggled to distinguish foreground objects from static backgrounds or ambiguous visual relationships without these complementary inputs.
Tried and failed
omitting digital elevation data in visual segmentation applied to aerial imagery semantic segmentation. Outcome: worse than baseline. Reason: lacks crucial topographic priors, causing false positive segmentations above natural elevation thresholds
Monitoring and understanding treeline dynamics in the Swiss Alps from 80 years of aerial imagery · EPFL · PhD
Tried and failed
omitting specialized high-contrast input modalities in segmentation applied to cortical lesion detection. Outcome: worse than baseline. Reason: standard input sequences lacked sufficient contrast to resolve subtle cortical boundaries without specialized imaging sequences
Deep learning-based analysis of multiple sclerosis lesions with high and ultra-high field MRI · EPFL · PhD
Tried and failed
random background sampling for semantic segmentation applied to satellite imagery land cover classification. Outcome: worse than baseline. Reason: underperformed compared to boundary-targeted and feature-based sampling strategies
Mapping agricultural land in support of opium monitoring in Afghanistan with Convolutional Neural Networks (CNNs). · Cranfield · PhD
Tried and failed
diffusion models applied to deterministic semantic segmentation. Outcome: worse than baseline. Reason: did not outperform standard discriminative models on deterministic prediction benchmarks
Fast and Future: Towards Efficient Forecasting in Video Semantic Segmentation · EPFL · PhD
Tried and failed
promptable foundation model post-processing for segmentation applied to color-sensitive semantic segmentation. Outcome: worse than baseline. Reason: expanded false positives by segmenting non-target objects sharing color features
A Machine Learning Approach to Recognize Environmental Features Associated with Social Factors · Virginia Tech · Master's
Considered and rejected
Considered and rejected: Single-scan batch inference for online segmentation was rejected due to lower accuracy from lacking historical viewpoint context.
3D Segmentation and Damage Analysis from Robotic Scans of Disaster Sites · Georgia Tech · PhD
Considered and rejected
Considered and rejected: Single-frame FPM semantic segmentation alone was rejected as a standalone BGS algorithm due to poor accuracy (0.5846 F1) when foreground classes appear static in the background.
Deep learning algorithms for background subtraction and people detection · Boston University · PhD
Tried and failed
fine-tuning 2D models for volumetric data applied to 3D medical image segmentation. Outcome: worse than baseline. Reason: lack of bidirectional volumetric context across slices
AI Systems for Understanding and Grounding Radiology Reports · Harvard · Master's
Considered and rejected
Considered and rejected: Rejected relying on manual slice/pixel ROI segmentation, as brain tumors vary widely in size/location and segmentation errors degrade classification while ignoring surrounding tissue context.
Deep-Learning Based Brain Tumor Classification · University at Buffalo · PhD
Tried and failed
pure vision transformer with smaller patch size applied to 3D spatiotemporal microstructural degradation prediction. Outcome: worse than baseline. Reason: Lacks inductive spatial biases of CNNs for complex 3D volumetric sequences
TransVNet: Predicting bone degradation using ViT and virtual dataset of cellular microstructures · Iowa State · Master's
Tried and failed
general-purpose self-supervised vision transformers applied to fine-grained visual skill assessment. Outcome: worse than baseline. Reason: recognition-optimized representations fail to capture subtle domain-specific articulatory quality nuances
Advancing Phonology-Based Sign Language Assessment: From Learner to Machine-Generated Videos · EPFL · PhD
Tried and failed
deep feature classification on unmasked raw images applied to pairwise visual disambiguation. Outcome: worse than baseline. Reason: pretrained visual features failed to distinguish ambiguous relationships without explicit foreground segmentation masks
Pushing the Boundaries of 3D Spatial Understanding · Cornell · PhD
Tried and failed
unimodal visual feature segmentation applied to spatial accident risk prediction. Outcome: worse than baseline. Reason: spatial dispersion and high false positive rates without auxiliary structural and dynamic context
Enhancing autonomous vehicle decision-making through scenario-based traffic rule integration · Imperial · Master's
Handcrafted intensity and histogram thresholding fails on complex foregrounds and variable lighting
Global, adaptive, and mixture model thresholding methods were unable to differentiate target objects from confounding background noise or hyper-intensities. These rigid heuristic segmentation strategies produced severe false negatives on low-intensity features and degraded downstream feature extraction.
Tried and failed
intensity thresholding segmentation applied to anatomical structure segmentation in medical imaging. Outcome: worse than baseline. Reason: confounding hyper-intensities prevent accurate separation of target structures from background tissue
Deep Learning for Localizing and Segmenting Anatomies in Medical Imaging · Cornell · PhD
Tried and failed
Gaussian mixture model automatic thresholding applied to time-lapse microscopy image segmentation. Outcome: worse than baseline. Reason: Failed to accurately segment features across time-lapse frames compared to a static threshold
Precipitation Dynamics at the Solution-Solution Interface in Confined Geometries, and the Effects of Organics on Precipitate Evolution · Georgia Tech · PhD
Tried and failed
color thresholding and skeletonization applied to overlapping root system segmentation. Reason: heuristic segmentation failed to separate complex overlapping structures from background
Tried and failed
simple intensity thresholding for segmentation applied to subcellular molecular cluster detection. Reason: missed low-intensity functional clusters and failed to resolve closely spaced assemblies
MECHANISMS OF TRANSCRIPTION FACTOR HUB FORMATION AND FUNCTION DURING EMBRYONIC DEVELOPMENT · Penn · PhD
Tried and failed
Automated intensity thresholding segmentation applied to subcellular focal adhesion quantification. Outcome: worse than baseline. Reason: Batch thresholding masks failed to detect features accurately, producing high false-negative rates versus manual quantification
Mechanotaxis and mechanoresistance in cancer · Imperial · PhD
Tried and failed
fixed-threshold normalized difference index segmentation applied to surface water mapping across regions. Outcome: did not generalise. Reason: a single threshold caused overestimation at some sites while completely omitting target features at others
Observing the Seasonal Evolution of Supraglacial Ponds in High Mountain Asia: A Supervised Classification Approach · Cambridge · Master's
Tried and failed
heuristic contrast adjustment and thresholding preprocessing applied to grayscale electron microscopy segmentation. Outcome: worse than baseline. Reason: degraded image quality and disrupted deep learning feature extraction
Deep Learning Approach for Cell Nuclear Pore Detection and Quantification over High Resolution 3D Data · Virginia Tech · Master's
Considered and rejected
Considered and rejected: Rejected global and adaptive threshold-based segmentation baselines (e.g., Otsu's method) because histogram-based thresholding inherently fails to distinguish root from non-root foreground noise.
ITErRoot: High Throughput Segmentation of 2-Dimensional Root System Architecture · Saskatchewan · Master's
Considered and rejected
Considered and rejected: Thresholding pixel values alone for laser line extraction (lacked repeatable precision required for structured light sensing, necessitating regional segmentation / binary image conversion).
Volumetric flow monitoring of biomass through an industrial grinder · Iowa State · Master's
Considered and rejected
Considered and rejected: Rejected using global thresholding segmentation due to poor multiclass classification compared to watershed and Bayesian/machine-learning approaches
Wettability characterisation of sandstone and carbonate rocks using X-ray micro-CT imaging · Imperial · PhD
Considered and rejected
Considered and rejected: Rejected purely automated thresholding (Otsu/RenyiEntropy) for final segmentation because overestimation and rigidity included background and cell bodies instead of true synaptic puncta.
Mapping of excitatory/inhibitory synaptic ratio in the lateral geniculate and mediodorsal thalamic nuclei · Boston University · Master's
Left open by the authors
Problems the authors named and did not get to.
Left open
Benchmark TEDM against foundation models like SAM for semi-supervised medical image segmentation under domain shift and out-of-distribution conditions. Blocker: None
Label-efficient medical image segmentation: the benefits and limitations of semi-supervised generative models · Imperial · Doctorate
Left open
Evaluate and improve unsupervised instance segmentation failure cases caused by background clutter and edge ambiguity. Blocker: None
Enhancing Unsupervised Instance Segmentation with Exemplars · Carleton University · Master's
Left open
Develop advanced post-processing techniques specifically tailored for gradient-based weakly-supervised semantic segmentation methods. Blocker: The thesis provides no specific design, algorithm, or mathematical formulation for the intended post-processing techniques.
Learning without Expert Labels for Multimodal Data · Virginia Tech · Doctorate
Left open
Implement and evaluate alternative classification algorithms beyond K-NN for historical document character classification in the press variant identification pipeline. Blocker: Lack of the specific historical print dataset and upstream segmentation pipeline output used in the thesis.
Automated identification of press variants in old documents · De Montfort University · PhD
Left open
Develop physiology-aware machine learning models to reduce segmentation errors in multiplexed tissue images. Blocker: Lacks specific definition, architecture, or formulation for what constitutes 'physiology-aware' models
Single-cell Methods and Spatial Analysis for Highly Multiplexed Tissue Images · Harvard · PhD
Left open
Develop generalist deep learning segmentation architectures beyond U-Net to segment multiple cellular compartments across imaging modalities. Blocker: Lacks specific architectural design requirements, target benchmark datasets, and evaluation metrics
Grounded: Inference via Local Signals & Learned Representations by Organic & Artificial Systems · Harvard · PhD
Left open
Extend weakly supervised satellite semantic segmentation to self-supervised frameworks with automatic error detection for refining low-resolution labels. Blocker: None
Weak-Supervised Deep Learning Methods for the Analysis of Multi-Source Satellite Remote Sensing Images · University of Trento · PhD
Left open
Combine pixel-wise, region-wise, and fuzzy boundary-wise loss functions and evaluate segmentation performance across standard metrics. Blocker: None
Incorporating fuzzy-based methods to deep learning models for semantic segmentation · Nottingham · PhD
Left open
Develop a transfer-learning module to generalize brain metastasis segmentation across varied institutional MRI acquisition protocols. Blocker: Access to diverse multi-institutional brain metastasis MRI datasets with expert segmentations
Advancing Radiotherapy Treatment Through Artificial Intelligence-Driven Approaches · UT Southwestern · PhD
Left open
Adapt the PrinCut-Auto segmentation framework to work across multiple imaging modalities such as MRI, two-photon, and confocal microscopy. Blocker: Lacks specific algorithmic techniques or concrete benchmarks for achieving generalizability across modalities
Animal Internal Motion Analysis with Unsupervised Machine Learning Methods · Virginia Tech · Doctorate
Who has worked on this
Authors of the public theses cited on this page, most-cited first.
- Mohammed Baharoon · Harvard · Master's · 2026
AI Systems for Understanding and Grounding Radiology Reports - Du Wang · Penn · PhD · 2026
INTEGRATION OF ARTIFICIAL INTELLIGENCE IN QUALITY ASSURANCE FOR HEAD AND NECK CANCER RADIOTHERAPY CLINICAL TRIALS - Margherita Rosnati · Imperial · Doctorate · 2024
Label-efficient medical image segmentation: the benefits and limitations of semi-supervised generative models - Jeff L Rhoades · Harvard · PhD · 2023
Grounded: Inference via Local Signals & Learned Representations by Organic & Artificial Systems - Mohammad D. Hossain · Queen's University · PhD · 2023
An improved segmentation and classification method for building extraction from RGB images using GEOBIA framework - Zeju Li · Imperial · PhD · 2022
Learning strategies for improving neural networks for image segmentation under class imbalance - Francesco La Rosa · EPFL · PhD · 2022
Deep learning-based analysis of multiple sclerosis lesions with high and ultra-high field MRI - Chen Chen · Imperial · PhD · 2021
Improving the domain generalization and robustness of neural networks for medical imaging - Jing Dao Dao Chen · Georgia Tech · PhD · 2021
3D Segmentation and Damage Analysis from Robotic Scans of Disaster Sites - Corey Zheng · Georgia Tech · PhD · 2026
Towards systems biophotonics in microscopy, medicine, and robotics - Luis Felipe Wolf Batista · Georgia Tech · PhD · 2026
Autonomous System for Identifying and Capturing Floating Waste - Hadi Shokati · Tübingen · PhD · 2026
A Unified Framework for Advancing Soil Erosion and Flood Assessment Through Deep Learning and Process-Based Modeling
Checking a claim in this area?
We can run the same search on any method or claim. If nothing turns up, we will say so, and that proves nothing on its own.