Abstract
Creating datasets for Neuromorphic Vision is a challenging task. A lack of available recordings from Neuromorphic Vision sensors means that data must typically be recorded specifically for dataset creation rather than collecting and labeling existing data. The task is further complicated by a desire to simultaneously provide traditional frame-based recordings to allow for direct comparison with traditional Computer Vision algorithms. Here we propose a method for converting existing Computer Vision static image datasets into Neuromorphic Vision datasets using an actuated pan-tilt camera platform. Moving the sensor rather than the scene or image is a more biologically realistic approach to sensing and eliminates timing artifacts introduced by monitor updates when simulating motion on a computer monitor. We present conversion of two popular image datasets (MNIST and Caltech101) which have played important roles in the development of Computer Vision, and we provide performance metrics on these datasets using spike-based recognition algorithms. This work contributes datasets for future use in the field, as well as results from spike-based algorithms against which future works can compare. Furthermore, by converting datasets already popular in Computer Vision, we enable more direct comparison with frame-based approaches.
Highlights
Benchmarks, challenges, and datasets have played an important role in the maturation of frame-based Computer Vision (Kotsiantis et al, 2006)
In this paper we focus on creation of Neuromorphic Vision datasets for object recognition
We have presented an automated process for converting existing static image datasets into Neuromorphic Vision datasets
Summary
Benchmarks, challenges, and datasets have played an important role in the maturation of frame-based Computer Vision (Kotsiantis et al, 2006). Quantitative evaluation of algorithms on common datasets and using common metrics allows for a fair and direct comparison between works. This ability to directly compare results encourages competition and motivates researchers by giving them a state-of-the-art target to beat. Datasets provide easy access to data for researchers, without which they would be required to gather and label their own data, which is a tedious and time-consuming task. A lack of publicly available Neuromorphic data means that Neuromorphic researchers must record their own data, which is in contrast to frame-based Computer Vision, where datasets can be constructed by assembling samples from an abundance of publicly accessible images. The barrier to acquiring Neuromorphic Vision sensors has recently been lowered significantly by Spiking Datasets for Neuromorphic Vision commercialization of sensors by iniLabs (Lichtsteiner et al, 2008), a lack of publicly available Neuromorphic Vision data and datasets persists
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
