Image quality plays a crucial role in vision-based safety-critical applications. From reliable dataset generation and training, to testing and evaluation, through to assisted and autonomous driving in adverse weather conditions, high-quality images are vital the accurate and reliable performance of the vehicles of the future. A lack of standardised KPIs has left the industry without a benchmark for performance evaluation, resulting in inconsistencies, hampering the industry’s ability to make objective decisions and improvements in this crucial area.

But, of course, balances throughout the vehicle and supply-chain such as cost, latency, and energy consumption must all be taken into consideration – how can we ensure the highest-quality images are available whilst maintaining these checks and balances?

In this panel, industry experts will discuss the impact of image quality on object, lane, and VRU detection, with a focus on challenging and complex light and weather combinations.