Quality

About Parameter

This parameter evaluates image quality based on specific criteria tailored for needs of reliable skin analysis, such as face positioning, lighting, resolution, face and skin visibility, occlusions.

Features

1. Works with 1 and 3 side facial images.

Output

1. score — overall numerical rating (0-100) indicating image quality. 0 means poor quality, 100 means excellent quality. Good quality images typically have a score > 70

2. Face and Skin Detection

   1. has_face — Boolean indicating if a face is detected in the image

   2. has_enough_skin — Boolean indicating whether enough skin area is visible for analysis (15% of the face-aligned image must contain skin)

   3. Head Position and Focus

      1. has_valid_pose — Boolean indicating if head pose is within acceptable parameters

      2. has_good_resolution — Boolean indicating if face resolution meets quality standards (at least one dimension exceeds 1024 pixels; for best results, we suggest to use images with at least 2K×2K face resolution to detect majority of visible pores)

      3. has_no_blur — Boolean indicating if image is free from motion blur

   4. Occlusion Detection

      1. has_no_occlusions — Boolean indicating if face is unobstructed

      2. has_no_eyeglasses — Boolean indicating absence of eyeglasses

      3. has_no_hair_bangs — Boolean indicating hair isn't covering forehead

   5. Lighting Quality

      1. has_good_exposure — Boolean indicating proper image exposure

      2. has_uniform_illumination — Boolean indicating even lighting across the face

      3. has_no_shadows — Boolean indicating absence of harsh shadows

---

Algorithm API

API access is available only for Professional plan clients