Research

Two peer-reviewed papers in computer vision and natural language processing, published under the name Owais Ansari.

01ADCIS 2025
Interpretable Instance Segmentation for Camouflaged Wildlife Detection
A segmentation framework for detecting camouflaged wildlife that pairs YOLOv8n-seg with explainable AI techniques — Seg-Grad-CAM and LIME — so that ecological monitoring systems produce not only accurate masks but auditable evidence for why a region was classified as an animal. Interpretability and segmentation accuracy are benchmarked against U-Net and FCN baselines.
- Best Presentation
- YOLOv8n-seg
- Seg-Grad-CAM
- LIME
- Instance segmentation
- Explainable AI
02ICT4SD 2025
AI-Based Synchronized Video Dubbing
A multilingual video dubbing system that chains automatic speech recognition, neural machine translation and text-to-speech synthesis while preserving audio–video timestamp alignment. The system supports ten Indian languages and runs on a hybrid backend: a Node.js API gateway in front of Flask services that host the AI inference workloads.
- Springer
- OpenAI Whisper
- Automatic speech recognition
- Neural machine translation
- Text-to-speech synthesis
- Timestamp synchronisation