Computer Vision Projects
Computer Vision Projects
Embedded Files
Keywords: Visual Search Engine, Video Understanding, Human Recognition
Human Search Engine - Examples
Human Search Engine - Examples
Animal Search Engine - Examples
Animal Search Engine - Examples
Plant Search Engine - Examples
Plant Search Engine - Examples
List of Projects:
List of Projects:
[1] Visual Search - Plant/Animal Search: training image recognition models on sub-billion scale noisily labelled images, to recognize tens of thousands of classes of plants/animals from input image. Challenges: noisy data, large number of classes, low storage/latency requirements, model acceleration, high intra-class different and inter-class similarity
[1] Visual Search - Plant/Animal Search: training image recognition models on sub-billion scale noisily labelled images, to recognize tens of thousands of classes of plants/animals from input image. Challenges: noisy data, large number of classes, low storage/latency requirements, model acceleration, high intra-class different and inter-class similarity
[2] Visual Search - Face Search: training face detection, face recognition, facial attribute recognition models for effective human recognition (celebrity search and similar face search). Challenges: training data availability, low storage/latency requirements for large scale real-time service, model inference acceleration (tensor RT and multi-process).
[2] Visual Search - Face Search: training face detection, face recognition, facial attribute recognition models for effective human recognition (celebrity search and similar face search). Challenges: training data availability, low storage/latency requirements for large scale real-time service, model inference acceleration (tensor RT and multi-process).
[3] Visual Search - Data Engineering: devising data cleaning models for large scale data cleaning for plant/animal/human datasets, with high recall/precision.
[3] Visual Search - Data Engineering: devising data cleaning models for large scale data cleaning for plant/animal/human datasets, with high recall/precision.
[4] Visual Search - Shopping Item Search: Recognizing clothes for their style and categories with high accuracy (>95%), applied on billion scale shopping SKU images in search engine backend dataset.
[4] Visual Search - Shopping Item Search: Recognizing clothes for their style and categories with high accuracy (>95%), applied on billion scale shopping SKU images in search engine backend dataset.
[5] Visual Search - Shopping Item Moderation: devised effective active learning approach to remove images with uncivil hand gestures.
[5] Visual Search - Shopping Item Moderation: devised effective active learning approach to remove images with uncivil hand gestures.
[6] Visual Search - Similar Image Search: Training large scale image representation model to allow similar image search. Key challenges: data collection and composition, low latency, open set problem, low latency and low storage requirements, evaluation ambiguities.
[6] Visual Search - Similar Image Search: Training large scale image representation model to allow similar image search. Key challenges: data collection and composition, low latency, open set problem, low latency and low storage requirements, evaluation ambiguities.
[7] VR-Glass hand gesture recognition: real-time hand gesture recognition on edge device, for interaction with VR glasses.
[7] VR-Glass hand gesture recognition: real-time hand gesture recognition on edge device, for interaction with VR glasses.
[8] Dense object instance segmentation in natural Scene images (AI Competition, Winner)
[8] Dense object instance segmentation in natural Scene images (AI Competition, Winner)
[9] Flash-based Face Anti-spoofing: Design and develop training dataset, AI models for flash based face anti-spoofing
[9] Flash-based Face Anti-spoofing: Design and develop training dataset, AI models for flash based face anti-spoofing
[10] AI based Audient Emotion and Gaze Analysis during TV-watching using ordinary RGB cameras.
[10] AI based Audient Emotion and Gaze Analysis during TV-watching using ordinary RGB cameras.
[11] Video Highlight Detection: detecting episodes in video for explosion/fighting events.
[11] Video Highlight Detection: detecting episodes in video for explosion/fighting events.
Page updated
Google Sites
Report abuse