CS543/ECE549 Past Project Topics
Fall 2022
- Small Object Detection Using Super Resolution Approach
- An end-to-end YOLOv5 detection and diagnosis system for breast cancer
- GAN-based Image to Image Translation
- Comprehensive Study on Monocular Vehicle Detection and Pose Estimation
- Classical face deblurring CNN models comparison and optimization
- Outfit Stylization and Cartoonization of People in Image Scenes
- Reimplementation of an Algorithm of Robust Lane Detection and Tracking
- Evaluation of the Robustness of Spatially-Adaptive Multilayer Selection for GAN Inversion and Editing
- Generating Dog Images Using Denoising Diffusion Models
- Reconstructing illumination and geometry for duplicated objects from few-view images
- Recognition and Reconstruction of Urban Scenes
- 3D Scene Reconstruction From 2D Images
- Real-time Structure from Motion with QR Detection
- Perception-Driven Human-Robot Chess Gameplay: Object Detection and 3D Pose Estimation for Autonomous Chess Piece Movement
- Face Mask Detection with YOLO-v5 and Faster R-CNN
- Feature Extraction from Nuclear Segmentation Maps for Breast Tumor Classification
- A Smart Robust Invisible Cloak in Multi-Scenario
- Auto-Scalor: Automatic Score Calculator for Board Games
- Automated Lane Segmentation and Environment Tracking
- Robust Visible Image Watermarking via Noisy Filtering
- Reimplementation and Investigation of MultiGrain: A Unified Image Embedding for Classes and Instances
- Extracting Pose from one Video and Reconstructing the Extracted Pose in another Video
- Exploring Visual Odometry Method using Stereo Camera/KITTI Datasets
- Real-world object recognition using privileged-information-based learning
- Knowledge Distillation and Mask R-CNN
- Efficient Video Semantic Segmentation with Only Information from the Compressed Bitstream
- Heatmap Generation of Soccer Player based on Motion Object Tracking and Localization
- Benchmarking the performance of ResNet, ViTs and efficient attentions
- Radon Transform: Another way to view 2D Structure with low-dimensional projections
- Explorations of Score Fusion Techniques for Visual Category Recognition
- Transfer of Human Motion from Video to a Virtual Humanoid Robot
- Lane Line Detection Dataset Expansion and nighttime lane line detection algorithm based on Cycle GAN
- An Investigation of NeRF Inpainting
- Human Re-Identification Based on Gait Recognition
- Boosting Fairness in Clinical Imaging Datasets via Generative Data Augmentation
- Comparison of Visual Question Answering systems
- Hand Tracking and Gesture Recognition for Automated VFX
- Facial Keypoints Detection using Deep Learning
- Kaggle Image Matching - Landmark Matching
- Sim-to-Real Vision-Based RL on Autonomous Vehicle
- Facial Expression Detection based on Different ImageNet
- Vehicle Speed Estimation using Single View Computer Vision
- KyoNeko: Vision-based Event Recognition of Household
- Novel Operators for Extracting Outlines of Buildings
- Image Segmentation and Object Detection for Chess Player Evaluation in a Live/Recorded game
- Text-based Web Captcha Cracker with Computer Vision
- Real-Time Partial Face Rendering/Reconstruction for Teleoperation
- Environment Mapping of Farmland Using Satellite Imagery
- Inpainting using Intermediate Layer Optimization in Generative Models
- Robustness of CheXNet Against Adversarial Attacks
- A Comparative Study on Metric Learning Techniques for Computer Vision
- Comparing and Improving Image classification using SIFT in SVMs and Convolutional Nets
- Recognizing Birds from Sound with Few–shot Learning
- Audio-guided Album Cover Art Generation
- Single-View Neural Geometric Level of Detail: Monocular 3D Reconstruction with Implicit Shapes
- Hand Recognition Exploration Using Deep Learning Methods for Computer Vision
- Tracking Objects as Points
- An Analysis and Comparison of FaceRecognition Techniques
- A Study of Automated Tumor Grading in Brain Cancer
- Violence detection with vision transformers
- An interface for generating, inverting, and editing card art with StyleGAN3
- Few-Shot Novel View Synthesis with Depth Supervision and Implicit Neural Representation
- Unsupervised Anomaly Detection using FastFlow
- Classifying Cloud Structures from Satellites
- Autopilot Driving - Traffic Sign Detection
- Reproducing Facial Feature Detection Using Geometrical Face Model: an Efficient Approach
- A Study of Inpainting with LaMa
- Is Unitary Scalarization still a Good Choice in Multi-Task Learning in Computer Vision?
- Image Super-Resolution with Deep Laplacian Pyramid Networks
- Exploring Different Spatiotemporal Models for Automatic Piano Transcription
- Diagnosis of COVID-19 Using Deep Learning on Chest X-Ray Scans
- Kaggle: Real-Time Facial Expression Recognition
- Image matching challenge: Local Features and Beyond
- Exploration of Super-Resolution with SRCNN
- Stepwise Goal-Driven Trajectory Prediction
- DreamDance: Personalized Text-to-video Generation by Combining Text-to-Image Synthesis and Motion Transfer
- Investigation of Style Mixing Methods on Multi-Domain Image-to-Image Translation
- Robust and Efficient Detector for Traffic Lights
- Carvana Dataset: Detection of Foreground Objects and General Segmentation Analysis
- A Comparative Study of Recent NeRF Models
- Comparison of Different Image Matching Approaches
- Landscape-GAN: Transforming the Landscape to Another with Generative Adversarial Network
- Exploring Transfer Learning in Face Mask Detection
Fall 2021
- Optical Character Recognition (OCR) from Images Enhanced with Super Resolution
- Predicting NFL player helmet impact from video and GPS data
- Object Detection in Complex Scenes using Lightweight DNNs
- Physics-based Learning of Parameterized Thermodynamics from Real-time Thermography
- FaceTime at a Cocktail Party: Towards High-quality Robust Audio-Visual Speech Extraction with Incremental Speaker Embedding
- Explainable deep learning-based classification on small-sized dataset
- Progressive Face Aging with Generative Adversarial Network
- Equivariant Point Set Analysis via Learning Orientations for Message Passing
- Trajectory Prediction of Heterogeneous Agents in Structured Scenes
- YOLOv5 For Simultaneous Classification and Detection on Difficult Datasets
- Segmentation of Face Attributes
- Novel Class Discovery Improves Few-Shot Detection
- Using Various Data Augmentations and Ensemble Techniques to Classify Faces Aging Images
- Offline Document detection and hand drawn Flowchart recognition with Faster R-CNN
- Kinship identification for pairs of faces
- Depth denoising in embodied AI
- Face Segmentation Models and Portrait Generation
- RF-GAN: GAN for Spoofing RF Heatmaps
- An examination of Faster R-CNN on the Object Detection Task in Regular and Hazy Environments
- Transfer Learning with segmentation using GAN in medical imaging
- Deep Visual Odometry Using Recurrent Convolutional Neural Network
- Noisy Policies for Visual Embodied Navigation
- Improving Neural Radiance Field with Integrated Positional Encoding
- Two-Stage Oriented Object Detectors
- Real time facial mask detection
- Visual Relation Modeling: Describing Relations between Images
- Lane Detection Based on Traditional & CNN Methods
- Multimodal Fine-tuning for Image Caption Retrieval
- Segment and label helmets in video footage
- Data Augmentation using GANs for Improving Brain Tumor Segmentation
- Optimization for Real-ESRGAN to Deal with Human Faces
- Horse to zebra translation
- Image copy detection using task-specific learnt image descriptors
- Motion Magnification with Eulerian and Learning-Based Methods
- Reimplementation of "Generating Classification Weights with GNN Denoising Autoencoders for Few-shot Learning"
- A General Approach: Object Detection with Region Proposals and Classification
- Human protein classification with transformers
- Hand-Drawn Graph Recognition
- A Comparison between U-Net based Lane Detection and Traditional Computer Vision based Lane Detection Method
- Estimating Water Absorption in Cement Pastes
- Classical and Neural Models for Multi-Frame Super-Resolution
- Unpaired image-to-image translation
- Manga style transfer
- Adversarial Frame Interpolation to Resolve Turbulent Flame Structures
- 3D Human Motion Prediction Using Transformer-based Models
- RANSAC-Flow: generic two-stage image alignment
- DA-DAE: Learning Good Image Embedding for Visual Agents via Data Augmented Denoised Autoencoder
- Examining and Refining Novel Implementations for Image Reconstruction
- Grading of Diabetic Retinopathy Severity Level
- Image captioning with attention in English and Chinese
- SIBAL: SNS Image Bias and Like Estimator
- Camera supervisor for Ring Fit Adventure
- Multimodal Recognition of City and Agroecosystem from Hyperspectral and Lidar Data: An Attention-Based Fusion Approach
- Object Detection and Segmentation on the COCO Dataset
- Smaller Vision Transformers for Performance on Datasets Small in Resolution and Size
- Interpreting Difference between Performance of Convolutional Neural Network on Classifying Bird Images
- Vehicle Tracking and Speed Estimation with YOLOv5 and DeepSORT
- Reproducing C2N: Practical Generative Noise Modeling for Real-World Denoising
Spring 2021
- Nuclei Instance Segmentation across Microscopy Imaging Experiments
- 2D Player tracking in ultimate Frisbee
- Segmentation and Classification of Brain Tumour from 3D Brain MRI
- Finding a way: Using UNet and Filtering to Find Routes Through Mazes and Campus
- Image and video resizing using seam carving
- Star detection with Deep CNNs
- Pulmonary Nodules Classification for Lung Cancer
- Weakly supervised learning for cancer diagnosis based on stain free multiphoton histopathology
- 3D aware Image composition with Poisson Blending
- Meta Learning for Continual Image Classification
- Generating Photo-Realistic Super Resolution Images with Generative Adversarial Networks
- Accuracy of Instance Aware Image Colorization vs Traditional Methods
- RPN for Faster RCNN on Multiple Models
- Using a GAN for data augmentation of chest X-Ray Images for detection and classification of COVID-19 virus in patients
- Improving Neural Art Style Transfer
- Optical Character Recognition for Traditional Chinese Handwriting
- PyTorchST: A library for neural style transfer
- HAODE-Net: Object detection in hazy environments
- Pedestrian Detection using Faster RCNN
- Image Style Migration
- Blob based object tracking
- Variation of Adabins on loss function, optimization and resolution
- Exploration of different methods of adversarial example attacks on deep neuron network and test their performance
- Implementing Efficient Deep Residual Network for Image De-noising
- 3D MRI brain tumor segmentation
- Implementation of YOLO Object Detector
- Self-supervised joint learning
- LiDAR-Tron: A Detectron2 immplementation for detecting cars using LiDAR point clouds.
- Towards efficient human robot collaboration using vision-based human motion predictions
- A Study of State-of-the-art Facial Expression Recognition Methods on Face Expression Recognition Dataset
- Evaluating the performance of SwaV on Chest Data
- Comparing semi-supervised learning methods for chest x-ray diagnosis
- Instance aware video colorization
- Image Translation from Photographed Images to Quick, Draw! Images
- Unsupervised deep clustering for skeleton based action recognition with Graph Neural Networks
- Optical Flow Based Moving Object Detection and Tracking
- Different approaches to face mask detection
- ResNet and DenseNet: Insights into Performance and Model Interpretability
- Multiple local cue based contour detection and how to break it with texture transfer
- Photo editing suite with focus on object removal and in-painting
- Visual Odometry using Stereo Vision
- Object detection from dynamic vision
- NeRF: Neural Radiance Fields for View Synthesis
- Learning based relative pose estimation for visual servoing of a soft robot
- RF-Turk: Scaling up dense image annotation in real world experiments with Inexpensive Robots and RFIDs
- Estimating the pose of small symmetric and reflective objects using only RGB Data
- Data augmentation using generative adversarial networks for segmenting OCT cornea images
- Faster RCNN Object Detection on Cars
- No-ball, Free Hit: Classifying Images and Comparing Accuracy for the Front Foot No-ball
- Cross-view localization using Monocular camera
- Amodal Video Instance Segmentation
- Super Resolution by Generative Adversarial Networks
Spring 2020
- An Evaluation of Generating Segmentation Masks for Autonomous Vehicle Application via Style Transfer Framework with Weakened Supervision
- Fashion Item Detection, Localization and Retrieval
- Human head tracking with MobileNet-SSD model in real-time video
- Analysis of Multi-frame Super Resolution Methods
- Video Object Tracking
- Find your lookalike
- Deep learning based Deraining Approach Comparison and Analysis
- Proximity Sensing using Monocular Camera
- Survey on Different Gaze Estimation Methods
- Shallow Water Specularity Filtering in Static and Dynamic Environments with Camera Motion
- Real-time Action Recognition in Livestock Surveillance feeds
- Dog Breeds Classification
- Automated Deception Detection in Videos
- Image Super-Resolution Based on Deep Learning Algorithms
- Visual Question Answering
- Automatic PCB Grader Project
- Bengali.AI Handwritten Grapheme Classification
- Monocular Depth Estimation on Autonomous Driving
- Agricultural Image Segmentation
- Fast Traffic Sign Detection and Recognition
- Lane and Stop Sign Detection for Autonomous Vehicles
- Comparison of Image Captioning With and Without Context Information
- Detecting Artifacts in DEM Images for Data Visualization
- YOLOv4 for Object Detection
- Are deep image super-resolution techniques robust to task-based evaluation in medical imaging?
- Imaging surface topography of pavements using photometric stereo
- Real Time YOLO Object Detection and Semantic Segmentation
- Neural Factorizations
- Online Super-Resolution using Deep Internal Learning
- Data Augmentation with SinGAN
- One-Shot Video Object Segmentation
- Universal Yolov3-SPP with GIoU Loss for Autonomous Driving Object Detection under Sunny and Foggy Weather
- Exploring Facial Recognition on iPhone with CoreML
- Sub-Pixel level Motion Tracking under Rain Weather Conditions
- Improving Neural Point-based Graphics
- FSM Objects From Images
- Attention-based Joint Detection of Object and Semantic Part
- 6D Pose Estimation
- Facial similarity recognition with Siamese Neural Network
- Wasserstein Style Transfer
- COCO Dataset: Object Detection and Segmentation Analysis
- Obstructive Artifacts Removal using Single-shot Images
- High Resolution Millimeter-Wave Vision For Self-Driving Cars: Extending to Multiple Cars
- Neural Style Transfer with Texture Networks
- Interpreting and quantifying deep representations
- Fetal Head Segmentation Using Ultrasound Images
- Re-unprocessing Images for Optimal CNN Denoising Performance
- GAN Controls from Its Latent Space
- 3D Object Detection and Mesh Generation
- Image-to-Image Translation Based on Wasserstein GAN
Spring 2019
- Generation of Semantic UI using Screenshot of an App
- Maximization of Mutual Information for Multimodal Image Registration
- Multi-Agent Diverse Multi-Video Summarization
- Cycle GAN implementation
- Reconstruct 3D model of a stadium
- Predict 3D object scene flow from binocular stereo
- Image Segmentation to Improve CycleGAN
- Classification on CIFAR-10 like dataset
- Photometric Stereo via Computer Screen Lighting for Real-time Surface Reconstruction
- OCR for Poker
- Changing Lighting/Shading for detecting forgeries
- Fashion Style Recognition
- Nondestructive Characterization Of Surface Cracks In Photovoltaic Silicon Wafers
- Diving into the Optical Flow
- Image Segmentation with FCN and DeepLab
- Making Spatio-Temporal Convolutional Neural Networks Scalable and Accurate
- DeepFashion Recognition
- Deepfake Faceswap
- Human Activity Detection
- Airbus Ship Detection Challenge
- Deep Learning Based Chinese CAPTCHA Cracking
- Face Detection and Emotion Identification
- Fuzzy logic for emotion recognition of facial expressions
- Comparison of different image matching methods
- Video Magnification and Audio Extraction from Visual Information
- Histopathologic Cancer Detection
- Detect defects in bridges using 3D point clouds and reconstruct
- Record Recognition App
- Workers Tracking in Reconstruction Environment
- Generation of faces of same person at different ages
- Manga Colorization
- Image Captioning
- Face Aging using Generative Adversarial Networks
- Generative Adversarial Networks for Balancing Biased Datasets
- Multiple Object Recognition
- Face and Keypoint Detection
- Dehazing Images Using Polarization
- Emotion Detection and Recognition
- Dog Breed Classification
- Face Swapping
- License Plate Recognition Systems
- Neural Style Transfer
- Image Inpainting using GANs
- Equivalent Source Reconstructionin RF Desense
- Interactive Photo Montage
- Image Inpainting
- DeepFashion image ranking
- Rate-control for Real-time Object Tracking
- 3D Character Animation from a Single Photo
- Deep Stroke-Based Rendering
- Object and Class Detector for Cars
- Image Matting
- Kaggle Whale Classification
- Fast Poisson Blending using Multi-Splines
- Automatically Locating Atom Defects In Electron Microscope Image by Deep Learning
- Face Aging with CycleGAN
- Segmentation of Microscopic Images
- Implementing Capsule Nets
- Sudoku Solver
- Traffic Sign Recognition
- Visual Based Real Time Tabulation for Guitarists
- Bag of Visual Words
- Video-Based Eye Tracker
- Study of Attention Mechanisms
- Investigating Generative Replay Methods for Image Classification in Continual Learning
- Motion estimation and dense optical flow prediction comparison
- Erlang system
- CycleGAN on Fashion Dataset
- Recognition in adverse weather conditions for autonomous driving
- Reconstruct the orientation of a car given an image
- Adversarial Methods in Traffic Sign Detection
- Time-Lapse Mining from Internet Photos
- Object Detection of Anime Characters
- Segmentation of clothes in Fashion Dataset
- 3D scene from 2D images
- MaskRCNN on another dataset
- Fast RCNN on Traffic
- Identify car's make and model
- Emotion Recognition
- Deepfakes generation
- Automatic Image Colorization
Spring 2018
- Translating Chinese paintings to photographs using CycleGAN
- Face editing
- Classifying Pavement Distress Types
- A Multi-Block-Matching Approach for Stereo
- Recovering spatial layout of a room with clutter
- Segmenting nuclei from cell images
- Real-time Rubik's Cube Solver
- Part-of-speech Guided Diverse Image Captioning
- Automated picture blending effects (neural style transfer)
- Regression-based Visual Question Answering
- Relation network for captioning
- Adversarial examples for image captioning
- Label-free part localization based fine grained recognition
- Recurrent Attention Convolutional Neural Network for Fine-Grained Image Recognition
- People counting from social media photographs
- Lane Line and Road Sign Detection
- Real-time detection of cars and humans, age classification of humans
- Image analogies
- 3D Voxel Patterns of Object Category Recognition
- Clothing segmentation with Mask-RCNN
- Kaggle Humpback Whale Identification Challenge
- Depth Map Prediction
- Phase-Based Video Motion Processing
- One-Shot Video Object Segmentation
- Google Landmark Recognition and Retrieval Challenges
- Adversarial sample generation using VAE
- Modeling semantic segmentation using deep structured methods
- Finding tiny faces
- Graph-based image segmentation
- Playing card recognizer and Blackjack player
- Feature detection and location-based matching
- Flower classification with CNNs
- Handwritten character and word recognition
- Neural style transfer
- Real-time face detection
- Fully Convolutional Networks for Semantic Segmentation
- Visual crack detection
- Long-term Recurrent Convolutional Networks for Captioning
- Handwriting recognition using multidimensional long short term memory
- Visualizing network response to adversarial examples
- Semi-supervised learning
- Real-time Multiple Object Counting and Tracking in Urban Mixed Traffic
- Labeling of brain structures in MRI images
- Concrete Slab Flatness Measurement using image-based Point Clouds
- Converting handwritten math to LaTeX
- A Comparison of Video Stabilization Techniques
- Traffic sign recognition
- Image super-resolution using GANs
- Single-view metrology
- Human key-points detection and 2D pose estimation
- Robustness of ORB-SLAM
- Handwritten music recognition
- Stylizing faces using portrait paintings
- Invasive Ductal Carcinoma Detector using GANs
- Understanding hand-drawn circuit diagrams
- Image to image translation using CycleGAN
- Handwritten math recognition
- Apparel keypoint detection
- Classifying Chinese seal script
- PacGAN to capture more modes
- Recognition of microorganisms in high-resolution nonlinear optical microscopy
- Automated diagnosis of melanoma
- Automatic product recognition
- Bringing portraits to life
- High Dynamic Range Imaging
- Detecting accessibility features in Google Street View
- Feature recognition on 2D floor plan images
- Classifying playing cards
- Image to image translation with GANs
- Contour-based object detection and tracking
- Improving image composition
- Boat Self-docking in GTA V
- CAPTCHA cracking
- Virtual keyboard
- Face swapping
- Balancing and Answering Binary Visual Questions
- Face detection and recognition
- 3D face reconstruction from a single image
- The Dodging Car: Teaching a robot car to dodge pedestrians
- Lagrangian Particle Tracking
- Show and Tell: A Neural Language Generator
- DenseNet for Kaggle furniture challenge
- CBIR for clothes retrieval
Spring 2017
- Image to Image Translation using cGANs
- Improving outlier detection for feature matching
- Inferring button press sequences from video game videos
- CNN with Cancer CT scans to determine if lesions are cancerous
- Line Art and Color from images
- Dense Depth Maps via Human-in-the-Loop Annotations
- Tool tracking in robotic surgery
- Ground plane segmentation for material identification
- Artistic Style Transfer
- Generative Modeling: GAN and DRAW
- State of the art object detector evaluation
- Image super-resolution
- Hand tracking and gesture recognition
- Neural Style Transfer
- Fan calculator: A scoring system for Mahjong
- Neural Algorithm of Artistic Style
- Tweaks to AI driving, Chatbots and Neural Style Transfer
- Vehicle Logo Recognition: Mercedes vs Audi
- Moving Objects Tracking in SLAM
- Fooling CNNs for Image Classification with Adversarial Examples
- Face detection using Viola-Jones algorithm
- Restoring Ancient Chinese Calligraphy Artworks using Computer Vision
- Image Recognition and Object Tracking
- Billiard detection and prediction
- Human Face Reconstruction
- Single image super-resolution
- Sea Lion Population Count
- 3D Face Reconstruction
- Gender classification
- Transfer Learning Based Food Recognition
- Lane Line and Road Sign Detection in Embedded System
- Playing cards recognition
- Image Captioning Based on Neural Network Approaches
- Indoor Localization Using Smartphone Images and WiFi APs
- Object Detection and Tracking in Grand Theft Auto V
- License Plate Detection
- NOAA - Stellar Sea Lion Population Count
- Cat Gender Recognition
- Automated I Spy
- Solderless Breadboard Debugger
- Pedestrian Recognition and Tracking
- Gender Classification
- Constructing Locally Dense Point Clouds
- Haar-based cascade classifier for face detection on embedded GPU platform
- Image generation from text descriptions using GANs
- Object tracking and inferring real world coordinates in basketball
- Vision-based human action recognition
Spring 2016
- Trajectory tracking and obstacle avoidance of a Quad-Rotor
- Monocular SLAM system
- Accurate positioning method for an Unmanned Aerial System
- Detecting Brain Tumor from multi-modal MRI using Convolutional Neural Networks
- OCR of comic books
- Printed Music Sheet Recognition
- An Implementation and Analysis of Picture-to-Identity Linking
- Identifying Game State via Images
- Identifying Cars by their Logo
- Composing Music Using Recursive Neural Network
- Real-Time Stereo Matching with Target Painting and GPU Acceleration
- Sympathetic Nervous Activation Estimation from Video Magnification
- Online visual-inertial odometry algorithm
- Dice reader
- Object detection with deep learning
- Virtual puppetry system
- Recognition and Volume Estimation of Food Intake using a Mobile Device
- Tracking trucks in videos (Caterpillar)
- Butterflies Species Detection Using Neural Network
- Fast R-CNN for object detection and action classification applied to excavators and dumptrucks
- Tagging of restaurant photos
- 3D reconstruction in Minecraft for real world
- Identifying Pedestrian Traffic and Behavior Patterns
- Image captioning
- License plate recognition
- Multimodal Deep Learning for Robust RGB-D Object Recognition
- Markerless Mobile Augmented Reality
- Extracting Vanishing Points across Multiple Views
- Vision-based structural health monitoring
- Non-rigid Face Tracking and Emotion Recognition
- Accelerated Trajectory detection for Real Time Video Processing
- Segmentation of cells in electron microscope images
- Real-time interactive Augmented Reality Chess Game Application
- Hold and volume identification in indoor climbing images
- Product price estimation from images
- Building occupancy tracking from video
- Vision Based 3-D Shape Reconstruction of Flexible Manipulators from Multiple Images
- Distributed Object Detection in Large Images
- Computer Vision Based Navigation of UAV to a Known Landing Zone
- Recognizing handwritten mathematical expressions
- Off-line recognition system for mathematical expressions
- Image Restoration and Low-level Vision via Sparse Coding and Deep Learning
- Count vehicles and to estimate the speeds of the detected vehicles from traffic surveillance camera videos in real time
- Facial recognition security system
- Machine Detection, Identification and Tracking (Caterpillar)
- Video Hyperlapse
- Deep Generative Model with partial input
- Truck tracking (Caterpillar)
- Car model classification, verification, and attribute prediction
- Real Time Tracking for Augmented Reality using OpenCV and Unreal Engine 4
Spring 2015
- Audio Editing: A Computer Vision Approach
- Reconstruct a time lapse of an arbitrary view point from point clouds generated from different time
- Semi-Automated Vertebrae Recognition in Flouroscopic Images
- Dice detection
- Determine safe landing zones
- View intepolation
- Integrating Augmented Reality and a SSVEP based Brain Computer Interface for Covert Face Recognition
- Spectral Clustering of Massive Image Dataset using Approximation Techniques
- ShopGuide
- Classifying different green infrastructure settings
- Simultaneous localization and mapping (SLAM)
- Using computer vision technology to detect drivers' alertness level
- RGBD object segmentation using object proposals
- Facial expression detection
- Human emotion statistics system using facial expression detection
- Object Discovery Using Density Based Clustering
- Efficient Reflection Separation from Videos
- Visual Font Recognition by Deep Networks
- Eye gaze tracking
- Image classification for Yelp
- Facial keypoints detection
- MCQer: The Automatic Grading Software
- Identify signs of diabetic retinopathy in eye images (Kaggle Competition)
- Landscape Time-lapse: day-to-day scene image stabilization
- Effect of shadow removal on object detection
- Straw Coverage Rate Recognition
- Recognizing art style
- Generating Sequences with Recurrent Neural Networks
- Optical Character Recognition
- Develop and use chromatags for localization
- Construction Equipment Pose Estimation via Convolutional Neural Networks
- GPU Speedups for Vision Algoritihms
- Monocular Depth Perception in Riverine Environments
- Multi-view object reconstruction
- Static Hand Gesture Recognition
Spring 2014
- Recognition and Localization of Traffic Signs in 3D Point Cloud Model
- Helping BAXTER robot see
- Measuring Room Dimensions using KinectFusion
- Real Time Augmented Reality Applications
- Video Google: Image Search in Videos
- Real Time Vanishing Point Estimation for Quadrotor Control in GPS-Denied Environments
- Motion Tracking and Optical Flow in a First-Person Shooter
- Classification of Galaxy Morphologies Using Support Vector Machines
- Computer Vision Based Real-time Storm Drain Detection
- An Effective Multi-View Synchronization Algorithm of a Moving Object
- Maze game level digitizer
- Optical music recognition
- Juggling coach
- Automated Identification of Mahjong Tiles
- Automatic sign and logo removal
- Deep Features for Image Retrieval and Clustering
- Visual Feedback by 3D Reconstruction of Elastic Rods for Robotic Manipulation Planning
- Dense 3D structure reconstruction of rowing boat
- Card tracker and scoring system for a card game
- Exploring Real-Time Face Detection in Portable Devices
- Crack detection on a Street Surface using a Cascade Object Detector
- Classifying images of green infrastructure
- Credit Card Reader with Auto crop and Transform
- Kinivacy: A Collection of Kinect Applications
- Eye gesture based interaction
- Thousand Aces: Identifying a Pinochle hand and making a corresponding bid
- Galaxy Classification with Deep Convolutional Neural Networks
- Integrating inertial-aided feature tracking and 3D reconstruction
- Extracting informative graphics from package designs
- Edge Detection Algorithm for Robot Guidance
- Structured Annotations for 2D to 3D Modeling
- Polysearch: Feature-based Image Search
- American Sign Language Detection Using Computer Vision and Supervised Learning Techniques
- Imaging of zebrafish
- Someone Like You: Similarity-based face search
- Hand gesture recognition
- Video Surveillance for Human Emotion Identification
- Body Rotation Detection
- Individual Tree Crown Delineation in High Spatial Resolution Multispectral Satellite Imagery
- Playing card recognition
Spring 2013
- Single-view metrology of rooms
- Identifying the gestures and motion of objects in video using a cell phone
- Face classification and biometrics
- Face detection with OpenCV
- Tumor segmentation in MRI volumes
- Font recognition
- Jigsaw puzzle solver
- Panorama stitching and face detection
- Carcassonne gameboard scorer
- Material classification for construction sites
- Activity analysis of construction video using RGBD images
- Thermal modeling of buildings
- Checking spacing of rebars in construction images
- Recognizing and solving sudoku puzzles
- Identifying playing cards
- Maze solving with a robot
- Motion stabilization of video sequences
- Image classification using Latent Dirichlet Allocation
- Interacting with virtual objects using a webcam
- Structure from motion from multiple viewpoints of a scene
- Recognition of hand gestures using Hidden Markov Models
- Hand detection and gesture recognition in video
- Obstacle avoidance for quadrotor
- Image quality prediction with saliency
- Matrix factorization
- Vehicle tracking and classification from a webcam
- Baseball trajectory extraction
- Smart video recorder
- Detection of workers on construction site
- Robust eye center detection for eye tracking
- BRDF estimation for computational relighting
- Tour into the picture
- Contour matching for 3D deformations
- PET system geometry calibration
- Augmented reality Android app
- Automatic creation of fake miniatures
- Blob detection on FPGA
Spring 2012
- Real-time object recognition in video games
- Multiple camera tracking with wide baseline homography
- Multi-view 3D object reconstruction
- Wild cat identification
- Background subtraction: comparison of three methods
- Taxi detection and tracking
- Classroom webcam control
- K-mer-based DNA sequence alignment using Inverse Document File approach
- Analysis of material composition
- Real-time segmentation of human data from Kinect
- Street sign recognition
- Control law and parameter recovery from video footage
- Aircraft localization and classification
- Detection of star patterns for space navigation
- Creation of 3D meshes using Kinect
- Automatic face tagging
- Object removal by retrieving background regions
- Segmenting out people in Kinect data
- Playing card recognition
- Depth upsampling using sparse dictionaries
- Monitoring and diagnostics of gas turbines
- Restoring motion-blurred images
- Power watershed segmentation
- Carcassonne gameboard scorer
- Image analysis of maize ear