Why Does Data-Driven Beat Theory-Driven Computer Vision?
- Resource Type
- Conference
- Authors
- Tsotsos, John; Kotseruba, Iuliia; Andreopoulos, Alexander; Wu, Yulong
- Source
- 2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW) ICCVW Computer Vision Workshop (ICCVW), 2019 IEEE/CVF International Conference on. :2057-2060 Oct, 2019
- Subject
- Computing and Processing
Cameras
Computer vision
Object detection
Lighting
ISO
ISO Standards
Training
dataset bias
sensor bias
deep learning
object detection
theory driven vision
- Language
- ISSN
- 2473-9944
This paper proposes that despite the success of deep learning methods in computer vision, the dominance we see would not have been possible by the methods of deep learning alone: the tacit change has been the evolution of empirical practice in computer vision. We demonstrate this by examining the distribution of sensor settings in vision datasets, only one potential dataset bias, and performance of both classic and deep learning algorithms under various camera settings. This reveals a strong mismatch between optimal performance ranges of theory-driven algorithms and sensor setting distributions in common vision datasets.