# Object Detection in Autonomous Vehicles: Status and Open Challenges

**Authors:** Abhishek Balasubramaniam, Sudeep Pasricha  
**arXiv:** [2201.07706](https://arxiv.org/abs/2201.07706)  
**Date:** 2022-01-19  
**Summary page:** https://www.summarizepaper.com/en/paper/2201.07706v1/

## Key points

- Object detection is a crucial computer vision task used in many consumer applications
- Autonomous driving heavily relies on object detection for safe and robust performance
- Deep learning-based object detectors are used to accurately determine objects in real-time
- Abhishek Balasubramaniam and Sudeep Pasricha discuss the state-of-the-art in object detectors for autonomous vehicles and open challenges for their integration into these vehicles
- Challenges include dealing with occlusions, different lighting conditions, accurately detecting small objects like bicycles or motorcycles, and handling unexpected scenarios
- Multi-modal sensor fusion techniques can enhance the accuracy of object detection models

## Summary

Object detection is a crucial computer vision task that has become an integral part of many consumer applications, including surveillance and security systems, mobile text recognition, and diagnosing diseases from MRI/CT scans. However, one of the most critical applications of object detection is in autonomous driving. Autonomous vehicles rely on the perception of their surroundings to ensure safe and robust driving performance. This perception system uses object detection algorithms to accurately determine objects such as pedestrians, vehicles, traffic signs, and barriers in the vehicle's vicinity. Deep learning-based object detectors play a vital role in finding and localizing these objects in real-time. In their article titled "Object Detection in Autonomous Vehicles: Status and Open Challenges," Abhishek Balasubramaniam and Sudeep Pasricha discuss the state-of-the-art in object detectors for autonomous vehicles and open challenges for their integration into these vehicles. The authors highlight that while significant progress has been made in recent years towards developing accurate object detection models for autonomous vehicles, there are still several challenges that need to be addressed. One major challenge is dealing with occlusions caused by other objects or weather conditions like fog or rain. Another challenge is ensuring that the object detector can operate efficiently under different lighting conditions such as nighttime driving or bright sunlight. Additionally, it is essential to develop models that can detect small objects like bicycles or motorcycles accurately. The authors also discuss the importance of designing robust object detectors that can handle unexpected scenarios like sudden changes in road conditions or unexpected obstacles. They suggest using multi-modal sensor fusion techniques to enhance the accuracy of object detection models further. Overall, this article provides valuable insights into the current state-of-the-art in object detectors for autonomous vehicles while highlighting some open challenges that need to be addressed before fully integrating them into these vehicles.