Partners
The VATICS M5S, leveraged advanced technology node 28nm, is a new-generation SoC designed for a variety of professional and consumer IP cameras including security cameras, home cameras, video doorbell camera, battery camera, 180° and 360°panoramic cameras, as well as home appliance and robot cameras. The M5S empowers a fast development of high-quality and multi-lens camera applications, offering upto 4-channels video-input, independent advanced multiple ISP pipelines (incl. multi-exposure High Dynamic Range fusion) with dedicated video quality enhancements for multiple wide-angle lenses & sensors, plus on-chip dewarping, stitching, and blending, all together for high quality real-time panoramic camera video processing. Besides, M5S permits significant CPU loading saving while keeping still a high flexibilities of distortion corrections required for various fisheye and wide-angle lenses to provide industry-leading quality of panoramic video reproduction.
The M5S integrates a 720MHz ARM® Cortex™ A5 CPU and NEON™, a CEVA DSP and rich peripherals for tailoring up every needs for a wide range of camera applications. The M5S provides high efficient video encoding for 5M@30fps HEVC(H.265) and 2M@60fps AVC(H.264) where newly developed SmartHEVC encoding can save up to 60% on network bandwidth and local storage when compared to standard H.264. It grants greatly improved accessibility to most of IP camera applications and a great deal of bandwidth and/or storage cost saving.
The M5S SDK provides Linux-based and RTOS-based framework and development environment that includes tuning tools and a rich set of APIs, enabling easy product customization and differentiation in areas such as sensor and lens tuning, audio and video analytics porting and refining.
VATICS keeps investing on advanced SoC process node, providing leading-edge SoC architecture design and implementation to achieve high performance and low power consumption.
- Combining powerful ARM Cortex™ CPU and CEVA DSP for massive computation
- Low power consumption enables the development of small form-factor and energy saving cameras
VATICS Smart HEVC Encoder makes significant reduction of useless or unimportant data wherever possible and maintains high-quality video without abrupting the network bandwidth.
- Featured with latest developments of Dynamic ROI, GOP, FPS and Noise Reduction for smartly detecting regions of interest especially the amount of motions in dynamic scenes; applying noise reduction with disparate strengths; and adjusting group of pictures and frame rate accordingly for achieving outstanding bit-rate saving without losing every important details
- Smart HEVC Encoder achieves extremely low bit rates to save up to 60% on network bandwidth and local storage costs compared to H.264 Encoder in video surveillance applications
- Combining powerful ARM Cortex™ CPU and CEVA DSP for massive computation
- Low power consumption enables the development of small form-factor and energy saving cameras
VATICS the 5th generation ISP is involved with the mainstream lens and sensor roadmaps to continue to pursue the fidelity of conversion of light captured by the sensor, so image and video can be recorded with perceptible details, authentic tonal layering, and lower noise.
- “Color Consistent” technology dynamically compensates to reduce color discontinuous and intensity
- “HDR” technology provides more discernible details in high-contrast scene and brings out ampler and accurate color presentation, making a more lifelike visual experience of the real world.
- “Smart Edge Enhancement” technology improves the image blurriness (out of focus) locally due to slightly misaligned of lens and sensor
- “Multi-Lens & Sensors ISP” technology compensates color and luminance deviations of video stitched and/or fused from sources of multiple lenses and sensors due to their production variations
VATICS 180°/360° multi-channel fisheye dewarping engine introduces an innovative way that incorporating a high-performance hardware-accelerated fisheye dewarpper, allowing real-time dewarping process for significant CPU loading saving, and flexible supporting geometry corrections for a diverse of fisheye/wide-angle lenses to provide industry-leading quality of panoramic video formation.
Dewarping
- Supporting up to 4-channel video-ins
- Algorithmic optimization for pixel/block-based transformations
Stitching and Blending
- Feature based matching
- Geometric model stitching
- Compensate color bias for different sensors
VATICS provides intelligent video analytics on-device (esp. on edge-device) so that permitting end-customers to have greater privacy protection, data locality and lower latency of data accesses in response to tasks operated on device site. VATICS “AI On-Device ” is NN-based design analyzing visual features in real time from live video, classified against a trained list of persons (esp. faces), pets, cars, and any objects of interest for target detections as well as recognition.
- Seamless tailoring up integrations with internal ISP and dewarping engines for high accuracy of detection & recognition rate even using wide-angle/fisheye lenses
VATICS creatively applies machine learning methodologies to enhance speech and audio signal processing. VATICS audio technology inspires novel and elegant solutions to those most challenging and longstanding technique difficulties in this field such as Noise suppression, Acoustic echo cancellation, Automatic gain control, Dynamic range control etc. Advantages include,
- Competitive performance with lower footprint. At most 30% computing resource reduction compared to solution on the market
- Lower hardware design limitation
- Machine learning based internal engine reduces enormous human tuning effort