공식 대리점

AI Computation Performance Improves and Rockchip Officially Launches Android 8.1 NNAPI SDK

2018-07-05

RK3399 features high performance, high scalability and versatility. CPU has applied the big.LITTLE architecture, Dual-core Cortex-A72 and Quad-core Cortex-A53, and its overall performance and power consumption have occupied a leading status technically. Meanwhile, GPU has applied the Quad-core ARM new-generation high-end image processor Mali-T860, and integrated much more bandwidth compression techniques to realize the excellent overall performance. In addition, Android 8.1 is able to completely invoke the neural neural network API (NNAPI) to realize the hardware acceleration, largely improving the AI Operational Performance of RK3399.

RK3399 SDK launched this time is a kind of GPU acceleration solution based on Android 8.1 NNAPI, which can provide corresponding acceleration support for related AI applications developed by means of Android NNAPI, and features the following four major advantages:

1. High Compatibility: Standard API can directly support various APK applications developed based on Android NNAPI;

2. Strong Universality: It can support numerous mainstream model architectures, and is suitable for various derivative and developed applications based on the mainstream model architecture, including face recognition, ADAS, goods identification and fatigue detection etc.;

3. Improved Performance: The real-time performance can be obtained among multiple tasks. For example, the highest frame rate can reach 23.2 when MobileNet is applied for image identification;

4. Lower Power Consumption: It can realize the high-efficient operation based on GPU, and the power consumption is only 1W when fully loaded

According to the APK Testing Data of image identification and target detection, the mainstream model’s performance is outstanding:

AI Computation has achieved an explosive growth, and Rockchip AI solution is widely applied to various industries including the image identification, the intelligent security, the intelligent drive, Automatic Speech Recognition, and the consumer electronics. However, Android 8.1 NNAPI Optimized SDK based on the RK3399 Platform is now officially launched to enhance the hard power of the AI solution Product Line.  

AI+VSLAM visual technology revolution, Rockchip helps in the industrial upgrading of the fourth-generation sweeping robots

Rockchip released the deep learning-based target detection solution, speeding up the commercial process of high-end AI

2018-05-17

On May 16, 2018, Rockchip released the deep learning-based target detection solution powered by RK3399. It can provide Turnkey solutions for high-end AI industry and support both Android and Linux with a target detection speed of more than 8 frames per second.

Target detection is a very popular research orientation in AI field. It involves the positioning and classification of target objects in pictures or videos. It is quite difficult for machines to acquire the abstract concept of objects and position them from RGB pixel matrix, which poses great challenges to the application of AI.

Currently, the major R&D orientations of AI technology includes: face detection, body detection, vehicle detection, QR code detection and gesture detection, which can be widely used in monitoring, intelligent transportation, new retails, natural interactions, etc., while the target detection technology is the basis of these applications. The deep learning-based target detection technology is highly accurate and robust, and yet cannot be actually deployed and applied in embedded devices for a long time due to heavy computation.

Upgraded to Android 8.1! Rockchip High Cost-effective Tablet Solutions Showed during HK Electronics Fair 2018

2018-04-13

Hong Kong Electronics Fair opened today, Rockchip announced that entry-level, medium and high - end tablet solutions fully support Android 8.1. RK3126C, RK3326, RK3368, RK3288 and RK3399 will better meet the diverse needs of the market. 

High cost-effective entry-level solution – RK3126C

RK3126C is the BEST choice of high cost-effective entry-level tablet solution in the market, which adaptable to Android 8.1 and supports Android Go GMS authentication.

• Quad-core Cortex-A7 CPU
• ARM Mali-400 MP2 GPU
• 1080P H.264/VC-1/MPEG/VP8 video decoder
• HD display

Mid-end tablet solution – RK3326

RK3326 is a new 64bit Quad-core mid-end tablet solution, which is extremely cost-effective and suitable for smart show products.

RK3326 adopts ARM 64bit Quad-core Cortex-A35 CPU; GPU is the latest ARM G31. It supports OpenGL ES 3.2, Vulkan 1.0; supports 32bit DDR4/DDR3/LPDDR3/LPDDR2 and so on, the configuration selection is flexible and diverse; abundant bandwidth design, which better supports higher resolution LCD and more complex application scene. It supports 8CH I2S/TDM/8ch PDM (suitable for smart show products).

High-end tablet solution – RK3368/RK3288/RK3399

High-end tablet solutions RK3368/RK3288/RK3399, can be applied to the high-end flagship tablet /2in1 tablet / education, finance, medical, game tablet and so on. Some series of Chromebook, Chrome tablet and other Chrome OS products are powered by RK3288/RK3399. 

RK3368—— Octa-core 64bit solution, Antutu score reaches 50,000

• Octa-core 64-bit Cortex-A53 CPU
• PowerVR G6110 GPU
• 32-bit DDR3/LPDDR3
• 1080P@60fps H265/H264/VP9 videodecoder
• FHD Display 

RK3288—— High performance Quad-core solution, Antutu score reaches 60,000

• Quad-core Cortex-A17 CPU
• ARM Mali-T760 MP4 GPU
• 64-bit Dual channel DDR3/LPDDR3
• 4K@60fps H265/H264/VP9 video decoder
• HDMI 2.0 output up to 4K60Hz
• Support Android/Linux OS

RK3399—— High performance flagship solution, Antutu score reaches 80,000

It adopts Dual-core Cortex-A72 and Quad-core Cortex-A53 Big.Little structure CPU, and Mali-T860 GPU. There are high performance, low consumption, rich interfaces and other characteristics. Significant features including:

• Integrated dual USB3.0 Type-C interface, Type-C Display Port audio and video output
• MIPI/eDP interface, supported 2560x1600 display and dual display
• Built-in PCI-e interface that supported high speed Wi-Fi and memory extension based on PCI-e
• Supported high bandwidth LPDDR4
• Dual ISP processing up to 13MP/S, supporting 3D images
• Compatible with Android, Linux, Chrome OS etc.

A thousand developers join the Rockchip Developer Conference

2018-04-20

The Rockchip Developer Conference 2018 is the third conference held by Rockchip in Fuzhou. The event attracted nearly one thousand developers and audiences from industrial leading companies and a number of renowned companies and partners with global influence attended the event.

A number of industrial leaders and pioneers, including Zhu Kaihua, Duer Business Division CTO, Xie Linfeng, the Chief Architect for IoT operating system of Alibaba Cloud IoT Business Division, Deng Qiuwei, the President of 360 Intelligence Business, Zhao Hengyi, the Vice President of AI Speech, Chen Feng, the Director of Product Management of ArcSoft, Gao Weihan, the CPO of Pisofttech and Song Xiangming, the Co-Founder of ReadSense gave wonderful keynote speeches in AI cutting-edge technology, AI interactive system, IoT, unmanned driving, major industries and major applications.

Rockchip breaks through in seven major areas

Rockchip demonstrated many cases of its cooperation with industrial giants at the conference. A number of chips of high performance and high cost performance have been successfully applied in nearly 100 industries, including the flagship chip RK3399, RK3288, RK3368 and RV1108.

In the application of high-end AI chips, the Cheetah AI Speaker powered by Rockchip intelligent audio solution newly released in March 2018 was demonstrated onsite. In addition to rich content, the AI speaker contains abundant built-in hot news on the Internet, music, reading and FM applications. It is greatly intelligent in man-machine interactions and able to quickly respond to questions such as “How’s the weather ?”, “Xiaobao, play a joyful song”, “Xiaobao, play a piece of news for today”, etc., which can satisfy user’s all-round needs for basic necessities of life.

Another demonstrated AI solution application was the SenseTime face recognition machine, which adopts the Rockchip image recognition technology and quickly recognizes faces for comparison with the ID card placed in the recognition area. The product may be gradually put into commercial use in the transportation sector nationwide.

In the application of mainstream and flagship-level chips, the cases of applications by the international giants Google and Sony were showed on site as well. Google has used Rockchip solutions for three successive generations on its electronic products for education, including the Chrome tablets and Chromebook notebooks. As Google Chrome OS is an online cloud system, teachers and students are able to retrieve all the personal data and information under their account by logging into their account on any terminal. Thus, it has strong application value and convenience in the education field.

Other exhibited items on site include the Sony reference design board ToF, which can be used in AR, VR and other fields that require automatic operation, such as robots and drones. Rockchip technology has been widely used in the base of the image technology.

Powered by Rockchip OP1-RK3399, Google released world’s first Chrome OS Chromebook Tab 10

2018-03-28

As defined by Google, the Chromebook Tab 10 is a lightweight version of Chrombook, designed for schools for convenient management and sharing. It is the first tablet that can operate Chrome OS, supports the superfine magnetic portable pen Wacom EMR stylus and is expected to support the Google Expeditions AR function. It can simultaneously support Chrome App and Android App. The product shape of Chrome OS+ tablet enables the interactive modes of touch and voice control to be more advantaged compared with keyboard.

In the design of hardware, the CPU adopts Rockchip OP1-RK3399, with a memory and storage space of 4GB+32GB, a 9.7-inch screen, a resolution of 2048x1536, a weight of only around 1.2 pounds, supporting Google Play and the USB-C extended function. It has an built- in HD 2MP front camera, double speakers, one microphone and a 5MP rear camera. The battery lasts up to 9 hours. The Wacom touch pen allows users to make use of the latest machine learning technology, bringing more natural handwriting experience.

Rockchip OP1-RK3399 was also used in the Chromebook plus jointly released by Google and Samsung at the CES in January 2017.
 

RK3399 CPU adopts Big.Little core structure, Dual Cortex-A72 big core + Quad Cortex-A53 little core structure. With Mali-T860 MP4 as the GPU, the general performance of the GPU improves by 45% compared to that of the previous generation. Additionally, the extensibility and multimedia performance of RK3399 are extremely strong, supporting DDR3\LPDDR3\LPDDR4 memory, H.264/H.265/VP9 4K10bit@60fps video decoding capability, HMDMI2.0, USB3.0, Type-C and even PCI-e port.

In recent years, Google has been making strong pushes towards the education market with steady increase in the market share. Chrome OS has gained a big success in the school market, with nearly all the computer manufacturers promoting the Chrome OS computers. By the third quarter of 2017, Chromebook had occupied 60% of the K-12 education market in the U.S.

The RK3399 OP1 Acer Chromebook Tab 10 will continue to assist in the upgrade and popularization of handwriting cloud tablet devices in the global education field.

Pulierde Linux Smart Cast A3 Officially Launched! Powered by Rockchip RK3036

2018-03-26

Pulierde has launched the Smart Cast A3. Being small, handy and easy to operate, the product is highly popular after the mass production and release into the market. The Pulierde A3 Smart Cast satisfies user’s demands for watching resources shared on big screens to a great extent. Thanks to the H.265 technology, users can play the H.265 encoded FHD videos online. Compared with the 720P image quality previously, it is a considerable enhancement in user’s visual experience.

Pulierde A3 adopts Rockchip RK3036 solution, which supports H.265 local hardware decoding, Android/IOS/Windows/MAC all-platform system, HDMI output and the transport Miracast/DLNA/Airplay Mirroring.

Three highlights of the RK3036 platform:

1. Save system space, eliminate redundancy and support the handy Linux OS

Compared with the traditional Android Miracast Dongle devices, RK3036 solution adopts Linux OS, avoiding the problems of stability and bad experience caused by mass redundant junk files in Android.

2. Save the cost of memory, demand for memory decreases by 10 times

The adoption of Linux OS RK3036 solution may reduce the cost of terminal products by 2/3. Compared to the reliance of the Android System on hardware, RK3036 supports hardware with a minimum memory of only 64M DDR+8MB FLASH, a decrease of nearly 10 times that of the demand of the Android System for memory. Thus, it has substantially optimized the cost of peripheral electronic components.

3. Support H.265 local decoding and Miracast experience of super low power consumption on mobile phones

It supports H.265 local hardware decoding for mirroring so that the mobile phone does not have to perform algorithm, thus substantially reducing the power consumption of the mobile phone. Compared to H.264, H.265 encoded videos save 50% of the bandwidth with the same image quality and the error rate is lower with the mosaic decreasing up to 59%. RK3036 is the only chip solution that supports H.265 among the entry-level dongle.

Phicomm released the first AI speaker R1 and created a top-quality sound experience together with Harman. It is the first time for Phicomm to step into the artificial intelligence filed.

The Phicomm R1 is powered by Rockchip RK3229, with quad-core Cortex-A7 CPU. It supports sound source localization, sound source enhancement, echo cancellation, and noise suppression technology. It is the first solution to support 8-channel digital silicon direct connection in the industry. It has 8GB storage space, which can store a large number of songs or nearly tens of thousands of minutes of recording, and supports 2.4G and 5G Wi-Fi frequency band.

The Phicomm R1 can recognize sound accurately, and the awakening rate is excellent under many scenarios. No matter in a noisy home, watching TV, or family members chatting loudly, the awakening rate of R1 exceeds 85%.

In the smart home ecology, Phicomm R1 will play the role of manipulation center. Based on the company’s internal smart product line, R1 can control intelligent platoon insert, query personal health information, and control smart home products such as air purifiers by voice.

The Phicomm AI Speaker R1 will gradually access more abundant functions and content and will realize more personalized services by AI technology, making life easier and more convenient. It will be open for sale on March 28, at RMB 2,499.

- More Rockchip AI solutions –

From February 27 to March 1, the 2018 embedded world exhibition & conference was held in Nuremberg, Germany. It is one of the world's most influential exhibitions and the world's largest embedded system exhibition.

In this exhibition, Rockchip showed its ultra-wide platform applications including embedded AI, industrial control, smart audio, smart vision and other development platforms.

A new AI product Sogou Travel Translator which is one of the important products of its AI strategy has been officially released by Sogou at the cooperative conference. Powered by Rockchip RK3399, the Sogou Travel Translator supports various translation modes such as text translation, dialogue translation, speech translation, real-view translation and provides translations in 17 languages. The Chinese-English offline translation capability enables users to use unhindered with non-networked.

RK3399 CPU adapts big.LITTLE core architecture with dual-core Cortex-A72+ quad-core Cortex-A53, which is technologically advanced in terms of overall performance and power consumption.

The Sogou Travel Translator based on RK3399 mainly solves the disadvantages of unstable signals and inaccurate translations of many translation Apps in the market. It combines highlights of the two new technologies and will meet the needs of travel better:

1. Offline translation technology, off-line neural network translation nmt and off-line OCR translation can be implemented in the terminal equipment to solve the stored problems and identify delay problems;

2. The array technology, dual microphone arrays, beam forming, and ambient noise reduction.It can find the strongest sound source, realize the noise reduction and achieve the highest level of speech recognition.

With the launch of Sogou Travel Translator, a strong member has added into nearly a hundred of industry applications of RK3399. RK3399 is suitable for hundreds of industry application products such as AI, electronic whiteboards, electronic schoolbags, face recognition equipment, unmanned aerial vehicles(UAV), robots, game terminals, gaming peripherals, mobile games on-hook servers, home appliances, advisement players/all-in-one machines, financial POS, onboard control Industry, thin client (cloud service), VOIP video conference system, educational tablet, karaoke entertainment, medical treatments, security / monitoring / policing, industrial control, IoT Internet of Things, VR shooting, VR etc.. It has features of high-performance, high-expansion, all-round application and it will empower AI and assists the layout of industry products.

The Next ODROID by Hardkernel.

ODROID-N1 key features:

- Rockchip AArch64 RK3399 Hexa-core processor
- Dual-core ARM Cortex-A72 2Ghz processor and Quad-core ARM Cortex-A53 1.5Ghz processor, big-LITTLE architecture
- Mali-T860MP4 GPU, support OpenGL ES1.1/2.0/3.0, OpenCL 1.2
- 4Gbyte DDR3-1866 RAM, Dual channel interface for 64bit data bus width
- 2 x SATA3 port, native SATA implementation via PCIe-gen2 to SATA3 interface
- eMMC 5.0 (HS400) Flash storage and a UHS capable micro-SD slot.
- 2 x USB 3.0 host port
- 2 x USB 2.0 host port.
- Gigabit Ethernet port
- HDMI 2.0 for 4K display
- 40-Pin GPIO port
- OS: Ubuntu 18.04 or Debian Stretch with Kernel 4.4 LTS, Android 7.1
- Size: 90 x 90 x 20 mm approx. (excluding cooler)
- Power: 12V/2A input (Attaching two 3.5inch HDD requires a 12V/4A PSU)
- Price: US$110 (To be adjusted based on DRAM market price changes)
- Mass production schedule: TBD

For more information about 'The Next ODROID' 

Please, refer to this website : https://forum.odroid.com/viewtopic.php?f=7&t=29932

RK3229 SOM is a cost-effective, optimized Turnkey Module with high performance that is certified by Google. It features Quad-core ARM Cortex-A7，up to 1.5GHz and integrates PMIC, Memory, eMMC, WIFI/Bluetooth with built-in fuel gauge. It is compatible with pre-certified Cast and the Google Assistant.

RK 3229 SOM is powered by Android Things to support peripheral hardware, such as speakers, motors, sensors and more. RK3229 SOM is optimized for IoT devices and there are several unique benefits:

1. It is a cost-effective SoM module certified  by Google for running Android Things with the Google Assistant . It can be used to for fast mass production of IoT devices and shorten the R&D cycle.

2. It is based on the Android Things platform. ODM/OEM/Developers can efficiently access Turnkey hardware/software and benefit product development and innovation.

3. Google services: pre-certified Cast and the Google Assistant in the module can greatly speed up development progress and improve stability.

4. It has built-in integrated software DSP, supports EQ/DRC audio, and is easier to develop and customize than hardware support.

5. It is based on Android Things, which benefits from regular security updates directly from Google.

6. It can deeply improve product updates and cost savings by Google OTA upgrade service.

It is easier for developers and OEM/ODM partners to launch products with it, since RK3229 SOM turnkey hardware solutions are available off the shelf.
 

RK3399Pro AI processor with superior general computing performance adopts big.LITTLE CPU, Dual-core Cortex-A72 and Quad-core Cortex-A53, showing technical leadership in whole performance and power consumption. The Quad-core ARM high-end Mali-T860 GPU which integrates more bandwidth compression techniques shows excellent overall performance.

RK3399Pro AI processor has high expansion capability, such as: supporting dual Type-C interfaces; dual ISPs  (each channel supporting a maximum 13 mega pixels); supporting 4096x2160 display output; supporting 8-channel digital microphone arrays input. With respect to software, RK3399Pro supports numerous APIs (Application Program Interface), including OpenGL ES 1.x/2.x/3.1/3.2, Vulkan 1.0, OpenCL 1.1/1.2, RenderScript etc.

RK3399Pro with super AI computing performance is Rockchip’s first AI processor by adopting CPU+GPU+NPU hardware structure. Its integrated NPU (Neural Network Processing Unit) incorporates Rockchip’s technology in fields of vision, voice processing and deep learning etc. Compared to traditional solution, the computing performance of typical deep neural network Inception V3, ResNet34 and VGG16 models on RK3399Pro is better and improved nearly hundred times.

Three important features of Rockchip RK3399Pro AI solution: 

l  High performance AI hardware
RK3399Pro adopted exclusive AI hardware design. Its NPU computing performance reaches 2.4TOPs, and indexes of both high performance and low consumption keep ahead: the performance is 150% higher than other same type NPU processor; the power consumption is less than 1%, comparing with other solutions adopting GPU as AI computing unit.

l  Superior platform compatibility
RK3399Pro NPU supports 8bit and 16bit and is compatible with various AI software frameworks. Existing AI interfaces support OpenVX and TensorFlow Lite/AndroidNN API; AI software tool s support the importing, mapping and optimizing of Caffe/TensorFlow model.

l  Easily development of turnkey solution 
Rockchip provides one-stop AI solution based on RK3399Pro, including hardware reference design and SDK. The solution can increase the AI products R&D speed of global developers and greatly reduce product launch time. It can significantly improve the speed of AI product development for global developers and greatly shorten time to market.

"The age of AI has come. As a global SoC manufacturer in China, we have the market layout of AI for many years. RK3399Pro is Rockchip’s first processor which integrates AI hardware. Its platform can be rapid MP for commercial." Chen Feng, the Global Vice President of Rockchip, said, "With super 2.4TOPs performance, low power consumption and abundant interfaces, this product is applicable to various AI application fields such as intelligent drive, image recognition, security monitoring, drones and voice recognition."

The release of Rockchip RK3399Pro has a great influence on global industry chain. It will greatly lower the threshold of innovation and mass production for commercial for AI industry, and accelerate the overall popularization of AI with higher performance and lower power consumption features.  

Integrating the ReadSense algorithm matrix, RV1108 has the front end super computing.

With the continuous development of AI in recent years, face recognition has become the mainstream technology trend. Through computer vision technology, it enables the terminal devices to have real-time intelligent computing ability is the potential factor for future outburst. Based on this, the updated version RV1108 with frontier ReadSense algorithm IP has the advantages of local computing ability and more market competitiveness.

RV1108 is an important product of Rockchip layout of AI and visual related fields. It has the key technologies of intelligent image processing: embedded with high performance CEVA XM4 visual processor DSP, up to 600MHz;8M ISP; Supports 1440p@30fps/1080p@60fps H.264 video codec; Built-in128M Byte DDR; High integration density; Supports MIPI DSI, HDMI TX, CVBS IN/OUT, 100M Ethernet, Audio Codec.

The powerful technology accumulation of the embedded underlying optimization of ReadSense, powers RV1108 withintelligent front-end computing power. In addition, the ReadSense optimized and customized CNN face detection for RV1108 CEVA XM4 processor, which has made full use of DSP to provide powerful computation ability. It has high robustness, and has good adaptability to light change, posture change, different skin colors, and the people of different ages.

Moreover, ReadSense algorithm IP based on RV1108 also has a completed algorithm matrix, which can perfectly calculate and run locally in the field of face detection, body detection, track, identify and so on, without relying on computer resources. It has extremely low power consumption, but the performance has enhanced several times. Such as face detection, RV1108 with ReadSense algorithm IP can reach 120ms detection speed and 90% detection rate at 1080P resolution. 

Core upgrading, aiming at "new" scene 
AI is rapidly moving towards long-tail scenes in industry applications. A new generation AI visual module based on Rockchip RV1108 can be flexibly used for face capture, new retail industry.Its advantage of local real-time computing will expedite industry product upgrading.

1. Smart camera

From the acquired image or video stream, it can analyze, detect and discern the facial images in real time and output the structured data, which can perform accurate face detection, recognition and alarm. In addition, the camera also traces and captures 10 people with the clearest faces in a proper angle. The capture is 99%, the false detection rate is less than 0.5%, and the capture repetitive rate is less than 10%.  

2. New retail and intelligent marketing

RV1108 AI visual module could realize passenger flow statistics, hot spot region statistics, membership recognition and other functions in the new retail industry application. In hotels, auto 4S shops and other places, it can realize intelligent marketing and service upgrading through intelligent management of front-end intelligent analysis and back-end cloud platform.  

Rockchip RV1108, which integrates powerful algorithmic, will have more possibilities in application scenarios, such as household IPC, smart cities and other fields. With the constant improvement of front-end intelligent perception and basic cognitive ability, more traditional fields will have a new wave of iterative development. Besides, the upcoming Rockchip RK1608 AI platform has powerful performance, which will push the new trend of technology in the field of AI chip. It is worth looking forward to. 

Upgraded Rockchip RK3328 UHD TV BOX Solution at Three Times the Speed

2017-10-16