Yogesh Gajula

Advanced Intelligent Systems & Deep Learning

Face Animation and Lip Synchronization Using Generative Adversarial Networks (GAN): This system generates lifelike facial expressions from a static image based on a provided audio file. Using GANs, the model analyzes phonetic cues from speech and creates synchronized facial movements, producing an animated video output. The project also supports cloud-based execution on Google Colab, allowing seamless file uploads, real-time progress tracking, and video playback with GPU acceleration.

Technologies Used: PyTorch, GANs, OpenCV, SpeechRecognition, Google Text-to-Speech (gTTS), Google Colab, CUDA, ffmpeg

     The process begins with audio processing, where speech-to-text extraction analyzes phonetic cues. Next, the GAN-based facial animation model generates realistic expressions synchronized with speech. Video compilation is then performed using OpenCV to create a seamless animated output. Finally, cloud deployment on Google Colab enables GPU-accelerated processing and efficient file handling.

This project showcases my expertise in deep learning, computer vision, and AI-driven animation, demonstrating how generative models can bring static images to life using audio input.

• Project Files
• Video

Computer Vision & Deep Learning

Eye Detection Using YOLO-NAS: This project implements a deep learning-based eye detection system using YOLO-NAS. The model is trained on a custom dataset to accurately detect eyes in images. The system supports real-time inference, enabling applications such as eye-tracking, gaze detection, and drowsiness monitoring. The project is integrated with Roboflow for dataset preparation and annotation, ensuring optimized training and evaluation.

Technologies Used: YOLO-NAS, PyTorch, Roboflow, OpenCV, Matplotlib, Google Colab, CUDA

     The project leverages Roboflow to annotate and convert the dataset into YOLO format. A YOLO-NAS model is trained for accurate eye detection, and the trained model performs real-time inference, overlaying bounding boxes on detected eyes. Performance is evaluated using training graphs and metrics, while Google Colab integration enables efficient, GPU-accelerated execution. 🚀

• Project Files
• Report

Virtual Reality Projects

Immersive Virtual Reality and Augmented Reality Development: I have developed immersive VR and AR applications featuring interactive gameplay, multiplayer networking, and real-time simulations using Unity 3D, Photon Engine, Vuforia, XR Plugin, and MOCK HMD.

Sea Dragon Game (Multiplayer): A real-time multiplayer game with combat mechanics and animated dragon actions. (Technologies: Photon Engine (PUN 2), Unity 3D, Particle Effects)
Roller Coaster Simulator (VR & Android): A VR simulation with physics-driven movement, immersive effects, and mobile VR optimization. (Technologies: Unity 3D, XR Plugin, MOCK HMD, Particle System, Physics Engine)
Augmented Reality Game: An AR-based interactive experience with real-time animations and AI-powered depth effects. (Technologies: Photon Engine (PUN 2), Vuforia Engine, Leipex AI, Unity 3D, Particle System)

     These projects highlight my expertise in VR/AR development, real-time rendering, physics-based simulations, and multiplayer functionality.🥽+🎮

• Project Files
• Files