Bricleir | Innovation Challenge: Synthetic Image and Label Data
Innovation Challenge: Synthetic Image and Label Data
Procter & Gamble
Closed: Submissions closed on 2/4/2021
Program/Special Project initiative

Synthetic Image and Label Data
Data enablers for Digital Product Innovation/Research

Business Background
P&G has been accelerating delivery of digitally enabled consumer products, some of which were show cased at CES over the last few years. Examples are; 

Opte -The Opte Wand scans the skin with an enhanced digital camera that captures 200 frames per second and instantly analyzes each image using a proprietary algorithm to detect tonal imperfections not visible to the human eye . 

Olay products – Olay Skin Advisor, Olay Labs and smart wand are all examples where consumers provide data via digital products to create a personalized beauty regimen. 

To create products such as these researchers require lot of different types both real and ‘synthetic’ data to understand and develop variety of real skin complexity including blemishes, aging signs, facial hair, skin tone that vary by gender. Acquiring this type of consumer research via traditional capture methods is a costly and time consuming and often evolves based on consumer needs 

Business Challenge: 

P&G uses images of consumers and products generated in research to develop new digitally enabled products. Evolving consumer needs requires more ‘data’ (e.g. image, video, text, audio, temporal signal ) to accelerate development of new digital products for consumers.
 We are looking for potential partners/providers to propose unique solutions, technologies to automate generation of ‘synthetic’ image and label data as a critical building block in developing algorithms and models. 

How can P&G apply the use of synthetic ‘Image’ & ‘label’ data to accelerate the development of digital products? 

ROI expectations

Relevant technologies and services may include, but are not limited to:
Solution that generates high quality images that have (almost) the same quality as real images. Images of interest can be;
• Consumer Products (packaged, unpackaged or in-use by Consumers)
• Consumers Faces of sufficient variety (e.g., gender, skin type, blemishes, aging features, head/facial hair styles, etc.)
• Images of objects before and after using specific products (e.g. Dirty/clean clothes, Teeth before and after using a
product, rash after using diaper/pad)

Areas Out of Scope
• We are NOT interested in low quality images
• Not Cartoony images that are distinguished from real images easily
• Not out of controlled variables (e.g. Only product X with {a,b,c} components)

The Ask:
If you have a solution to partner with us, please submit a summary of your solution, approach and/or technologies that can help empower “Synthetic Image and Label Data” via our open innovation portal in section
Please indicate how you would propose pilot testing your solution. We look forward to partnering with you. All ideas and submissions will be fully reviewed. Please note that only non-confidential information can be accepted for review.

You must log in to respond to this initiative.
Know of another company that can help Procter & Gamble with this initiative?

Community Comments

Be the first to comment: