TetrasMobile Video
Image SDK Solution

Tetras.AI's intelligent imaging software solution delves deep into the entire process of phone photography, from sensors to ISP, to the entire application. It is an end-to-end computational photography solution with a product line that covers multiple key scenarios such as night scenes, super resolution, portraits, and videos, assisting phone manufacturers to comprehensively improve photo/video imaging quality and enhance users' photographing/video shooting experience.

Home > Products & Services > TetrasMobile Intelligent Image SDK Solution > Night Sight
Product Introduction

Imaging capabilities in low light and night scenes have always been one of the core selling points of mobile phones, and a pain point for users. Due to exposure and hardware limitations, imaging effects under dark light are generally poor. Our night scene photography product solutions cover full-scene and full-focal-length use. These multi-frame Raw Domain solutions support handheld multi-exposure when the mobile phone's exposure is insufficient, solving image quality problems with brightness/contrast/dynamic range/color/clarity/detail/noise under dark light, while also providing the ultimate performance experience. These solutions allow night scenes to be presented in their true beauty.

Tetras.AI's night scene photography products include our Super Night, Fast Night, and Night Portrait, along with other functional products for a variety of scenes such as dark light, extremely dark light, and dark light portraits. This way, one set of SDK solutions can flexibly support all or some product functions. Superior user experience is achieved by maximizing the use of the platform's computing power through heterogeneous computing and load balancing and scheduling. Our self-developed "multi-platform self-adaption" and "tuning" system ensures that the SDK can be adaptively matched with new platform drivers, and has efficient debugging and delivery capabilities under high concurrency and multi-project pressure.

Our night photography products employ the latest image processing and AI technology to process continuous multi-frame images received from the phone's front and rear cameras. They are also supported by multi-frame noise reduction, AI noise reduction, 3-4 stops of dynamic exposure fusion, dynamic exposure frame number and sequence, portrait enhancement and other new technologies to optimize each image quality dimension in dark light. We provide customers with full process delivery services, from sensor-adaptive exposure-control output, RAW domain processing algorithms, ISP debugging, to patch integration.

Product Advantages

The advantages of our night photography solutions include:

1. Product scope: covers the full-scene, full-focal-length, full-module, and various modes of the RAW domain. One set of SDK solutions can flexibly realize all or some product functions, including Super Night, Fast Night, Night Portrait, HDR, and MFNR functions.

2. Delivery capacity: our self-developed "multi-platform self-adaption" and "tuning" systems ensure that the SDK adaptively matches new platform drivers and has efficient debugging and delivery capabilities under high concurrency and multi-project pressure.

3. User experience: pursues a superior user experience while enhancing image quality. The computing power of different platforms is fully utilized through heterogeneous computing and load balancing, and scheduling. The solutions also support multi-frame streaming processing, further improving the experience of time-intensive photography.

4. Superior image quality: various usage scenarios are customized and optimized with the latest image processing and AI technology, resulting in the most appropriate brightness while retaining the ambiance of the night scene. Backlit portraits are bright, vibrant, and saturated with no overflow. The dynamic range is more hierarchical without noticeable or severe ghosting of moving objects. Pictures are clean and clear, and portraits present rich details and natural skin color. Support image quality enhancement in extremely dark 0.1 lux scenes.

Key Technologies
  • AI Exposure

    3-4 stops of dynamic exposure fusion, and dynamic exposure frame number and sequence are used to make scenes bright and transparent, especially backlit portrait scenes. Robust highlight suppression and dark detail restoration functions are also provided; the dynamic range is more layered and eliminates ghosting caused by moving objects.

  • AI Noise Reduction

    Multi-frame noise reduction and AI noise reduction are used to judge the level and form of noise in different areas of the picture to remove noise (luminance noise and color noise) while retaining sufficient detail and texture, balancing noise and detail.

  • AI Image Quality Enhancement

    Varying degrees of exposure and inter-frame redundancy information are employed, and high resolution is achieved through AI image quality enhancement to restore even more picture detail, resulting in picture quality that is pure and clear.

  • AI Portrait Enhancement

    Face verification and AI portrait enhancement technology are used to make facial brightness natural, light and shadow uniform, clearly defining facial features and making skin color natural in portrait photos.

Application Scenarios
  • Supports dark light, extremely dark light, dark light portraits, backlit portraits, moving portraits and other scenes
  • Supports front and back main cameras, wide angle, and other modules across the full range
  • Supports Night Scene Mode, Normal Mode, Portrait Mode, High Pixel Mode, etc.
×
Other products
Cooperation
Partner with Tetras.AI to Co-create New Possibilities