2D Boxing Annotation and Image Classification in the E-Commerce Industry
Image classification and labeling projects are common in the e-commerce industry.
The main disadvantage of online shopping is that consumers cannot directly “feel” the products, and images are the primary source of product perception, especially regarding clothing combinations and accessories. We classify and mark a large number of images so as to provide consumers with accurate results and make the search more efficient.
We’d like to share a 2D image classification and annotation project in the e-commerce industry.
1 All objects that appear in an image should be labeled with boxes.
2 Each box contains only one object and matches the outline of the subject
- The makeup here mainly refers to cosmetic products for the face.
- Cosmetics means products like shower gel, shampoo, toothpaste, and wash supplier, excluding lipstick and other beauty makeup.
- Accessory here mainly refers to something people wear but doesn’t include glasses, hats, and belts.
- Glasses include ordinary glasses and sunglasses
- Appliance (electrical appliances) here mainly refers to mobile phones, TV, and air conditioners, excluding rice cookers, fans, juicers, shavers, and Roomba.
- The food here mainly refers to steak, sausage, etc., not including something of infant formula/vitamin.
- Wine here mainly refers to wine, liquor, etc., excluding alcoholic beverages of beer.
- Soft drink mainly refers to non-alcoholic beverages, such as mineral water, juice, etc.
- For gender, it is essential to label after clearly identifying the gender of and seeing the face and body of the person in the image.
- Other refers to targets that are not listed in the 35 categories.
How it works on ByteBridge Dashboard
You configure and we annotate MANUALLY.
Data Export-By Yourself Service
1 For dense objects, it is necessary to label them one by one, such as a dozen of lipsticks.
2 The jewelry that the model is wearing should be labeled, such as a necklace, ring, and bracelet. Please zoom in on the image and label them.
3 If there are two shoes in the image, it is necessary to label them with two boxes.
4 If there is an occlusion, only the visible part should be annotated.
5 Sleeves should be included in the top. The shoes and the top should be excluded from the bottom.
6 The box needs to cover the entire shoe, including the toe and heel.
Outsource your data labeling tasks to ByteBridge, you can get high-quality ML training datasets cheaper and faster!
- Free Trial Without Credit Card: you can get your sample result in a fast turnaround, check the output, and give feedback directly to our project manager.
- 100% Human Validated
- Transparent & Standard Pricing: clear pricing is available(labor cost included)