Vehicle Object Detection in the context of Bangladesh Traffic

Sust-DLenigma Kaggle Competition Solution

• Field : Computer Vision

---

Dataset

The dataset covers the following 9 districts in Bangladesh: Sylhet, Dhaka, Rajshahi, Mymensingh, Maowa, Chittagong, Sirajganj, Sherpur, and Khulna. This dataset contains 9825 images with 78,943 Objects with 13 different classes. It is annotated with rectangular bounding box.Participants will encounter a wide range of road types, including towns, expressways, highways, and village roads. This diversity in locations aims to challenge algorithms to perform well across various driving contexts commonly encountered on Bangladesh roads.

• Train Image Count: 5896
• Test Image Count: 1964

---

Models

We employed transfer learning using three different approaches.

• YOLO configurations, ranging from YOLOv6L6 to the advanced YOLOv8
• Novel transformer-based models such as rtDETR and CoDETR.
• Faster R-CNN.

---

Authors

• Ayesha Binte Mostofa (1805062)
• Sumonta Nandy Amit (1805069)
• Md. Asif Haider (1805112)
• Mashiyat Mahjabin Prapty (1805117)