Budget: 1500 USD Deadline: 150 days
Hello to you.
For such systems, one of the initial stages is the training of recognition.
The training goes on the original data, i.e. Photo + characteristics.
The more, the better.
For example, a copy of the description of the data set for shoe recognition:
Data Understanding
Class Distribution: Train Set - 10,000 images (2,000 per class), Test Set - 1,215 images (Variable distribution across classes), Validation Set - 2,500 images (500 per class).
Pixel Value Distribution: 0 ~ 255 pixels.
Size Distribution: As depicted in the chart.
Data Quality: No improper data types or low-resolution images were detected.
