An Automotive Case Study on the Limits of Approximation for Object Detection
Authors:
Martà Caro,
Hamid Tabani,
Jaume Abella,
Francesc Moll,
Enric Morancho,
Ramon Canal,
Josep Altet,
Antonio Calomarde,
Francisco J. Cazorla,
Antonio Rubio,
Pau Fontova,
Jordi Fornt
Abstract:
The accuracy of camera-based object detection (CBOD) built upon deep learning is often evaluated against the real objects in frames only. However, such simplistic evaluation ignores the fact that many unimportant objects are small, distant, or background, and hence, their misdetections have less impact than those for closer, larger, and foreground objects in domains such as autonomous driving. Mor…
▽ More
The accuracy of camera-based object detection (CBOD) built upon deep learning is often evaluated against the real objects in frames only. However, such simplistic evaluation ignores the fact that many unimportant objects are small, distant, or background, and hence, their misdetections have less impact than those for closer, larger, and foreground objects in domains such as autonomous driving. Moreover, sporadic misdetections are irrelevant since confidence on detections is typically averaged across consecutive frames, and detection devices (e.g. cameras, LiDARs) are often redundant, thus providing fault tolerance.
This paper exploits such intrinsic fault tolerance of the CBOD process, and assesses in an automotive case study to what extent CBOD can tolerate approximation coming from multiple sources such as lower precision arithmetic, approximate arithmetic units, and even random faults due to, for instance, low voltage operation. We show that the accuracy impact of those sources of approximation is within 1% of the baseline even when considering the three approximate domains simultaneously, and hence, multiple sources of approximation can be exploited to build highly efficient accelerators for CBOD in cars.
△ Less
Submitted 13 April, 2023;
originally announced April 2023.
