{"id":179727,"date":"2025-04-25T09:07:00","date_gmt":"2025-04-25T07:07:00","guid":{"rendered":"https:\/\/www.aivancity.ai\/blog\/?p=179727"},"modified":"2025-04-30T09:09:46","modified_gmt":"2025-04-30T07:09:46","slug":"from-coursework-to-cutting-edge-deep-learning-innovations-in-classroom-by-graduate-students-at-aivancity","status":"publish","type":"post","link":"https:\/\/aivancity.ai\/blog\/from-coursework-to-cutting-edge-deep-learning-innovations-in-classroom-by-graduate-students-at-aivancity\/","title":{"rendered":"From Coursework to Cutting-Edge: Deep Learning Innovations in classroom by Graduate Students at aivancity"},"content":{"rendered":"\n

By Anuradha Kar, <\/strong><\/strong>Associate Professor of AI & Robotics at aivancity<\/p>\n\n\n\n

Introduction<\/h2>\n\n\n\n

As the domain of artificial intelligence continues to evolve at a rapid pace, it\u2019s not just researchers in labs or company engineers who are pushing boundaries\u2014young students also play a vital role in driving AI innovation by bringing fresh perspectives, curiosity, and bold experimentation. When engaged in hands-on projects and research, they challenge conventional methods, explore untapped ideas, and often pioneer creative solutions. Their passion and adaptability make them key contributors to shaping the future of artificial intelligence.<\/p>\n\n\n\n

This year, I integrated hands-on projects into my Deep Learning for Computer Vision (DLCV) course for the first-year Master\u2019s cohort (equivalent to the fourth year of the Grande \u00c9cole programme). After working through a semester, the students delivered an outstanding range of projects that were both technically ambitious and socially meaningful, with strong ties to real-world applications.<\/p>\n\n\n\n

What makes their works even more commendable is that these projects were conceived and completed by the students within the short timeframe of continuous assessment (as we call them \u201ccontr\u00f4le continu\u201d) as part of the DLCV course for this semester. Each student proposed their own project idea and developed it independently within a period of three months\u2014guided by iterative discussions and feedback which I included as part of my DLCV lectures. My role as a mentor was to advise, challenge, and support their thinking while ensuring their original ideas remained at the heart of their work.<\/p>\n\n\n\n

Below I share some of the details and outcomes of these projects, highlighting the students\u2019 creativity, technical skills, and their ability to apply deep learning to meaningful, real-world challenges.<\/p>\n\n\n\n

From ideas to impact: Student project spotlights<\/h2>\n\n\n\n

Here\u2019s a glimpse into some of the key project themes and brief descriptions of student projects that emerged across the cohort. The project implementations can be found in the linked GitHub repositories.<\/p>\n\n\n\n

Theme 1: Innovations in Visual Learning for Human Representation<\/h3>\n\n\n\n

In response to the growing need for detecting AI-generated content, one of our students, Abdellahi El Moustapha developed DualSight, a multi-task image classification system that combines object recognition and authenticity verification. Leveraging YOLO-based pseudo-labeling for object detection and a ResNet-50 backbone for shared feature extraction, the model was trained on a large dataset of approximately 152,000 images, achieving high accuracy (85\u201389%) in distinguishing real from synthetic images. An interactive demo is deployed on HuggingFace, with future plans to refine labeling strategies and improve detection performance. Project repository: https:\/\/github.com\/Abmstpha\/dual-sight-project<\/a><\/p>\n\n\n

\n
\"\"
Fig 1. A sample result from the deep fake detector framework named DualSight showing accurate classification of a deep fake (but realistic looking) image as \u201cAI Generated\u201d<\/figcaption><\/figure>\n<\/div>\n\n\n

Another student Ahmed Ben Aissa\u2019s project addresses a common e-commerce challenge\u2014predicting clothing sizes based on user-submitted images. His two-stage system first uses a modified EfficientNet model to estimate body measurements from front and side views. Then, a CatBoost Regressor maps these measurements to standard clothing sizes. The models are trained on BodyM dataset (1) and e-commerce data. Through this project we could identify limitations in available open source clothing datasets for computer vision (e.g lack of clothing type specific size features) and potentials to develop brand-specific models, augmented reality integration, and improved data privacy. Project repository: https:\/\/github.com\/zaizou1003\/Predicting_Sizes_CV<\/a><\/p>\n\n\n\n

\n
\n
\"\"
(a)<\/figcaption><\/figure>\n<\/div><\/div>\n\n\n\n
\n
\"\"
(b)<\/figcaption><\/figure>\n<\/div>\n<\/div>\n\n\n\n


Fig 2(a)\u00a0 Features of the BodyM dataset with body measurement profiles (b) Face landmark detection project by Darryl: green points: ground truth landmarks. red points: predicted landmark<\/p>\n\n\n\n

Other projects in this category include Oussama Bouriga’s real-time emotion recognition system using a custom CNN trained on FER-2013 dataset to detect five emotions from real time webcam video input (https:\/\/github.com\/Oussamabouriga\/emotion_detection<\/a> ) . Darryl Towa developed a facial landmark detection model based on a modified ResNet18 that predicts 68 landmark points from grayscale images, trained on the iBUG 300-W (2) dataset. Baptiste Langlois’s project aims to enhance construction site safety using a YOLOv5-based model trained on Roboflow dataset to detect helmets and heads.<\/p>\n\n\n\n

Theme 2: Activity recognition in complex and dynamic environments<\/h3>\n\n\n\n

This category attracted some remarkable projects from students such as the one by Likhita Yerra which focuses on detecting and analyzing chess moves from video using computer vision and NLP. It uses YOLOv8 to identify all 12 chess pieces on an 8×8 board, tracks movements, and converts them into standard chess move notation. Based on these, a GPT-4 LLM is used to generate game commentary for users to understand chess moves . It is deployed as an interactive Streamlit app where users can upload videos for running the pipeline<\/p>\n\n\n\n

. She named the project \u201cchess vision narrator\u201d and the implementation may be found at https:\/\/github.com\/LikhitaYerra\/Chess-Vision-Narrator<\/a> Future plans include real-time processing, dataset expansion, integration with Stockfish (3) for move evaluation, and multi-platform deployment. A video demo showcases the pipeline in action, highlighting the blend of technical accuracy and engaging commentary.<\/p>\n\n\n\n

\"\"
Fig 3 Summary of the chess vision narrator<\/em><\/strong> workflow from the presentation of Likhita Yerra<\/figcaption><\/figure>\n\n\n\n
\n