Skip to content

DS flows

In the table below, we have detailed the tech tools we leverage to identify and take down fakes, with examples of how these are deployed to better identify HUL fakes.

Table 61
Column 1
Column 2
Column 3
Image
Detection Tool
Process
Outcome
Text on Image
The model will highlight certain texts which include brand names anywhere in a particular image. The results will then be shared for manual QC for analysis. Based on such analysis, Meesho team verifies whether the results are genuine cases of counterfeits Accordingly, the genuine cases are processed for deactivation.
Our tool has flagged the word ‘Rin’ in the image. Since, the word is not evidencing any counterfeit or attempt thereof, this product has passed the manual QC process and continues as ‘active’.
rin.png
Logo Detection
The model will scan the images for logos (whether existing on the platform or manually fed) by running thorough checks on the images. The results are then shared with QC team for manual analysis Upon independent manual investigation, the results which fail to pass manual QC are processed for de-activation
Since we noticed a number of escalations from HUL regarding infringements of its 9 to 5 brand, the same was input into the tool which detected the above image when the logo ‘9to5’. Based on the flag, the product in question was delisted.
Logo detection.png
Image Match
The model conducts an exhaustive scan basis curated images and historical internal data from previous deactivations; Once the scan is complete, images are verified through manual QC and accordingly processed for deactivation.
The image match tool detected similarity between the above product and previously delisted copies of Lakme Eyeconic, despite the fact that the brand name i.e. ‘Lakme’ is not visible on the product.
3rd.png
Keyword based Flagging
The model capability is expanded to scan text as well for matching key words appearing in product description.
The seller attempted to list a product using the Lux keyword, which was flagged by this model. This helped identify indirect counterfeiting and the product was deactivated.
4th.png
Compliance Gating
Create a gating mechanism at the time of catalog upload to add a ‘checker mechanism’. This process blocks the listing from going live for the sellers without brand authorization.
The gating process displays a ‘failure message’ in case the seller tries to list a brand that is protected under our Suraksha List (i.e. an internal categorization created for well known brands prone to counterfeiting where we deploy stringent checks to ensure that only authorized suppliers are permitted to list products of such brands). This helps us filter out the attempted counterfeit listings.
5th.png
There are no rows in this table
Want to print your doc?
This is not the way.
Try clicking the ··· in the right corner or using a keyboard shortcut (
CtrlP
) instead.