Training Domain Specific Models for Energy-Efficient Object Detection

Kentaro Yoshioka, Edward Lee, Mark Horowitz

Abstract: We propose an end-to-end framework for training domain specific models (DSMs) to obtain both high accuracy and computational efficiency for object detection tasks. DSMs are trained with distillation and focus on achieving high accuracy at a limited domain (e.g. fixed view of an intersection). We argue that DSMs can capture essential features well even with a small model size, enabling higher accuracy and efficiency than traditional techniques. In addition, we improve the training efficiency by reducing the dataset size by culling easy to classify images from the training set. For the limited domain, we observed that compact DSMs significantly surpass the accuracy of COCO trained models of the same size. By training on a compact dataset, we show that with an accuracy drop of only 3.6%, the training time can be reduced by 93%.

TL;DR: High object-detection accuracy can be obtained by training domain specific compact models and the training can be very short.

Keywords: Object detection, efficient, domain specific models