You are here

AutoSeg: Automated Image Segmentation & Classification

Award Information
Agency: Department of Defense
Branch: Army
Contract: W911QX-21-P-0056
Agency Tracking Number: A202-107-0228
Amount: $111,487.03
Phase: Phase I
Program: SBIR
Solicitation Topic Code: A20-107
Solicitation Number: 20.2
Timeline
Solicitation Year: 2020
Award Year: 2021
Award Start Date (Proposal Award Date): 2020-10-16
Award End Date (Contract End Date): 2021-07-29
Small Business Information
1221 Brickell Ave., Suite 900
Miami, FL 33131-0000
United States
DUNS: 079640621
HUBZone Owned: No
Woman Owned: No
Socially and Economically Disadvantaged: No
Principal Investigator
 Stanislav Shalunov
 (415) 601-7021
 stas@clostra.com
Business Contact
 Gregory Thiele
Phone: (415) 377-8051
Email: gthiele@clostra.com
Research Institution
N/A
Abstract

Recent innovations in deep learning theory and implementation have enabled neural nets to achieve what was once unthinkable:  beat humans at complex image recognition skills, safely pilot cars over chaotic road systems, and overwhelm Grandmaster Lee Sedol in the game of Go, a challenge previously thought immune to AI because of the game’s near-infinite complexity. Clostra has applied a machine learning solution to automatically label/segment ML datasets.  While training neural networks are computationally intensive and requires specialized hardware,  execution is computationally inexpensive and can be implemented with very modest CPU and memory requirements.  Phase 1 of the project determines feasibility by testing and training a sophisticated segmentation/localization/classifier allowing for large datasets to be automatically label for use in training neural networks.

* Information listed above is at the time of submission. *

US Flag An Official Website of the United States Government