ai · 2020-06-18 · • vegetable recognition engine • ingredients recognition engine...

3
AI

Upload: others

Post on 09-Jul-2020

3 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: AI · 2020-06-18 · • Vegetable recognition engine • Ingredients recognition engine (chopped/mixed vegetables) • Drink recognition • Test environment setup inline with end

AI

Page 2: AI · 2020-06-18 · • Vegetable recognition engine • Ingredients recognition engine (chopped/mixed vegetables) • Drink recognition • Test environment setup inline with end

© 2020 TATA ELXSI. ALL RIGHTS RESERVED.

A I e n g i n e A c c u r a c y

95% for all AI modules16 recipes recognized

M a r k e t

Japan

D e l i v e r y M o d e l

ODC

• Vegetable recognition engine• Ingredients recognition engine (chopped/mixed vegetables)• Drink recognition• Test environment setup inline with end product environment

Microwave oven which recognize the food ingredients kept in it, and sets up the cooking time and temperature

B e n e f i t

Easy operationReduced cooking time

EngineeringCustom AI network

P r o d u c t p h a s e

In R&D

S C O P E

C H A L L E N G E S• Image pre-processing to make it suitable for specific local market

requirements• Visibility issue, poor lighting, camera position• Porting and optimization for embedded platform• Custom Architecture development

• TensorFlow• Custom designed network

T O O L S A N D T E C H N O L O G I E S

U S I N G D E E P L E A R N I N G TO E N H A N C E C O O K I N G E X P E R I E N C E

Page 3: AI · 2020-06-18 · • Vegetable recognition engine • Ingredients recognition engine (chopped/mixed vegetables) • Drink recognition • Test environment setup inline with end

© 2020 TATA ELXSI. ALL RIGHTS RESERVED.

A I e n g i n e A c c u r a c y

90%

D e l i v e r y M o d e l

ODC

• Machine vision for object recognition (obstacle detection), optimized cleaning, benchmarking of algorithm

• Detection and avoidance of wires, carpets, fallen objects, pedestal legs• Person avoidance• Data set generation• Mapping, path planning and navigation• Test environment setup inline with end product environment

Intelligent robotic vacuum cleaner that can detect and avoid obstacles

M a r k e t

Japan

EngineeringFunctional prototype

P r o d u c t p h a s e

In testing

S C O P E

C H A L L E N G E S• Detection performance improvement from 0.2 fps to 2.2 fps on

Qualcomm board• Meeting memory and performance requirements • Data set collection, cleansing, and segregation• Optimization of algorithm to run on an embedded system

• OpenCL• TensorFlow

T O O L S A N D T E C H N O L O G I E S

A I B A S E D N AV I G AT I O N F O R O P T I M I Z E D C L E A N I N G

Field validation

*Image for representation purpose