top of page
arm Mali.png

Arm Mali Image Signal Processors: C32, C52, C55, C71AE, C78AE & C720AE

Modern embedded vision systems rely on image signal processors (ISPs) to convert raw sensor data into usable images for displays, computer vision, and AI applications. The ISP sits between the camera sensor and the rest of the processing system, performing functions such as noise reduction, HDR processing, tone mapping, color processing, and image enhancement.

Arm's Mali Camera family includes several ISP solutions designed for different embedded and automotive applications. The Mali-C32, Mali-C52, Mali-C55, Mali-C71AE, Mali-C78AE, and Mali-C720AE provide different combinations of image quality, camera support, performance, power efficiency, and functional safety.

​​​

What Is an ISP?

An image signal processor processes data coming directly from an image sensor and transforms it into useful image outputs.

A simplified camera pipeline is:

Image Sensor → ISP → RGB / YUV / Other Output → Computer Vision / Display / AI

The ISP can perform many operations, including:

  • Auto exposure (AE)

  • Auto white balance (AWB)

  • Noise reduction

  • HDR processing

  • Demosaicing

  • Color correction

  • Tone mapping

  • Sharpening

  • Lens and shading correction

  • Image scaling and processing

The exact capabilities depend on the ISP architecture and configuration.

Arm's Mali ISP family is designed for applications ranging from embedded and IoT cameras to automotive ADAS and machine vision.

​​​

Mali-C32

The Mali-C32 is an area-optimized configuration of the Mali-C52 architecture. It is intended for low-power and cost-sensitive embedded vision applications where the complete feature set of a larger ISP may not be required.

The ISP supports a single sensor interface and can support up to four cameras. It supports image resolutions up to 16 megapixels and provides HDR, spatial and temporal noise reduction, tone mapping, and RGB/YUV output capabilities.

Typical applications can include compact embedded cameras and other products where silicon area, power, and cost are important design considerations.

​​​

Mali-C52

The Mali-C52 provides higher-performance image processing for embedded applications where image quality and dynamic range are important.

It supports high dynamic range processing and can be optimized for either image quality or silicon area. Arm specifies support for up to four cameras and resolutions up to 16 megapixels.

The Mali-C52 is designed for applications such as security cameras, smart-home devices, drones, and other intelligent imaging systems.

One important aspect of the C52 is its broader ISP software solution. Arm provides ISP drivers, 3A libraries for auto exposure, auto white balance, and auto focus, as well as calibration and tuning tools.

​​​

Mali-C55

The Mali-C55 moves toward higher-resolution, energy-efficient processing for embedded and IoT vision applications.

It supports multiple camera sources and can process up to 48-megapixel images, with throughput up to 1.2 gigapixels per second. Arm positions the C55 for smart cameras, commercial and consumer security, drones, robots, AR devices, and other edge-AI applications.

A key advantage of the C55 is its integration with machine-learning accelerators. Processed image data can be passed directly toward an ML accelerator, allowing AI processing to become part of the camera pipeline.

The C55 also provides HDR capabilities and configurable implementations that can be optimized for image quality or silicon area.

​​​

Mali-C71AE

The Mali-C71AE is designed for higher-performance automotive and industrial imaging systems.

Unlike the earlier embedded-focused ISPs, the C71AE places significant emphasis on functional safety and multi-camera computer vision. It supports up to four real-time cameras or up to 16 camera streams from memory and can process up to 1.2 gigapixels per second.

The C71AE supports several sensor color filter array configurations, including RGGB, RCCC, RCCB, RCCG, and RGBIr. This flexibility is useful for different automotive and machine-vision sensor architectures.

Applications include:

  • ADAS

  • Lane detection

  • Collision avoidance

  • Industrial inspection

  • Mobile robotics

  • Mirror replacement

  • Night vision

The C71AE also includes built-in features intended for functional-safety applications, with Arm providing a safety package supporting requirements associated with ISO 26262 and IEC 61508.

​​​

Mali-C78AE

The Mali-C78AE builds on the C71AE architecture while incorporating image-quality capabilities from Arm's embedded ISP products.

It is designed primarily for automotive systems where cameras may need to support both machine vision and human-facing displays. The C78AE can support up to four real-time camera inputs or up to 16 camera streams from memory.

This makes the C78AE particularly relevant for systems that need to process several cameras simultaneously.

For example, an automotive platform could use camera inputs for surround view, ADAS, driver assistance, or other perception functions while also generating high-quality images for displays.

The C78AE also provides functional-safety capabilities intended for automotive and industrial applications.

​​​

Mali-C720AE

The Mali-C720AE is Arm's newer automotive ISP designed for advanced driver-assistance systems, digital cockpits, and software-defined vehicles.

One of its defining features is a dual computer-vision and human-vision pipeline. This allows the ISP to process image data differently for machine perception and human viewing rather than forcing both applications to use the same image-processing configuration.

The C720AE supports up to 16 sensors, HDR processing through a native 24-bit pipeline, and multiple color filter array configurations including RGGB, RCCB, RCCG, RYYCy, and RGBIr 4x4. Arm specifies throughput of up to 1.5 gigapixels per second.

Another important feature is its differentiable ISP model. This can be used to accelerate image-quality tuning and help optimize ISP processing for computer-vision workloads.

The C720AE also provides multiple parallel outputs for functions such as scaling, cropping, foveated vision, and Gaussian pyramid generation.

​​​

Choosing Between the Mali ISPs

The different Mali Camera products are intended for different system requirements.

The Mali-C32 is focused on area and cost-sensitive embedded systems.

The Mali-C52 provides a strong combination of image quality, HDR, and configurable processing for embedded cameras.

The Mali-C55 increases performance and resolution while maintaining an emphasis on energy efficiency and edge-AI applications.

The Mali-C71AE introduces higher-performance multi-camera processing and functional-safety capabilities for automotive and industrial systems.

The Mali-C78AE builds on that architecture for multi-camera automotive applications where computer vision and display processing need to coexist.

The Mali-C720AE extends the architecture further with dual computer/human vision pipelines, advanced automotive processing, and support for newer vision workloads.

The right choice depends on factors such as camera count, resolution, frame rate, HDR requirements, sensor type, AI workload, power consumption, silicon area, safety requirements, and the desired image-quality pipeline.​

​

ISP Tuning and Camera Integration

Selecting an ISP is only one part of developing a camera system. The ISP must also be configured and tuned for the specific image sensor, lens, lighting conditions, and application.

A typical integration process may include:

Image Sensor → Sensor Driver → Camera Interface → Mali ISP → ISP Tuning → Image Output → CV / AI Application

ISP tuning can involve parameters for exposure, white balance, color correction, noise reduction, HDR, tone mapping, sharpening, and other image-processing functions.

The sensor and optics also need to be evaluated together with the ISP. A high-performance ISP cannot compensate for every limitation introduced by an unsuitable sensor, lens, or optical design.

Arm provides ISP software, calibration, and tuning tools for its Mali Camera products, helping developers configure the ISP for specific sensors and image-quality requirements.​

​

Mali ISPs in Embedded Vision Systems

The ISP plays an important role in determining the quality and usefulness of camera data before it reaches computer vision or AI software.

For example:

Camera Sensor → Mali ISP → YUV / RGB → Preprocessing → NPU / GPU → Computer Vision → Application

In an automotive system, multiple cameras may instead feed the ISP simultaneously:

Multiple Cameras → Mali ISP → Vision Processing → ADAS / Display / AI

This makes ISP architecture an important consideration when designing embedded vision hardware and software.

​​​

Conclusion

Arm's Mali Camera family provides ISP solutions across a broad range of embedded and automotive applications.

Mali-C32 and C52 target embedded imaging, while Mali-C55 increases performance and camera resolution for energy-efficient edge-AI systems. Mali-C71AE and C78AE extend the platform into high-performance, multi-camera automotive and industrial applications, while Mali-C720AE introduces newer capabilities such as dual human/computer-vision pipelines and differentiable ISP modeling.

For embedded camera development, the ISP must be considered together with the image sensor, optics, camera drivers, hardware platform, image-quality tuning, and computer-vision software.

​​​

For embedded camera development, the ISP must be considered together with the image sensor, optics, camera drivers, hardware platform, image-quality tuning, and computer-vision software. Explore our hardware design services to learn how we support custom camera integration, ISP development and tuning, schematic and PCB design, and complete embedded vision hardware development for systems using Arm Mali Camera ISPs.

bottom of page