Multilayer Output Buffer for Image Processing Aberration Correction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Image processing systems face challenges with distortion aberration and chromatic aberration, particularly in wide-angle lenses, which require costly optical systems or inefficient image processing techniques, leading to frame delays, power consumption issues, and limited data transfer speed.

Innovation Solution

An image processing apparatus that employs a multilayered output buffer configuration with a buffer controller adding tag information including camera IDs and pixel addresses, enabling efficient distortion correction and chromatic aberration correction through time division processing and burst transfer, reducing buffer size and improving transmission efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If image processing is performed through a DRAM, then image processing capability is improved, but frame delay occurs and data transfer speed is limited due to DRAM bandwidth

Engineering Contradiction:
Improveimage processing capabilityVSAvoidframe delay
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

The patent divides the output buffer into multiple layers (first layer, second layer, third layer) to segment the buffering process. This allows different stages of image processing data to be handled in parallel across layers, reducing the time bottleneck that would occur in a single-layer sequential processing through DRAM.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces a temporal dimension by implementing time-division multiplexing where the single image processing circuit processes multiple image pickup units at different time slots. The multilayer buffer structure provides additional spatial dimensions for data organization, allowing efficient data staging without requiring proportional increases in DRAM bandwidth.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Productivity

If image processing is performed through a DRAM, then image processing capability is improved, but power consumption increases

Engineering Contradiction:
Improveimage processing capabilityVSAvoidpower consumption
Core Design Contradiction:
ProductivityVSUse of energy by moving object

Solution Approach 1:

By segmenting the buffer into multiple layers, the patent reduces the frequency of data transfers to and from DRAM. Each layer can be filled and emptied independently, allowing the system to batch operations and reduce the overall number of high-power DRAM access cycles, thereby lowering power consumption while maintaining processing capability.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The multilayer output buffer acts as an intermediary between the image processing circuit and the DRAM. It allows data to be staged and prepared in the buffer layers before being transferred to DRAM in optimized batches, reducing the total number of DRAM access operations and consequently reducing power consumption.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Device complexity

If multiple image pickup units are processed with one image processing circuit in time division manner, then device complexity is reduced, but processing speed for each unit decreases

Engineering Contradiction:
Improvenumber of image processing circuitsVSAvoidprocessing speed per image pickup unit
Core Design Contradiction:
Device complexityVSSpeed

Solution Approach 1:

The patent implements preliminary action by pre-processing and staging data in the multilayer output buffer before final DRAM transfer. The first layer receives and preliminarily processes data from the image processing circuit, preparing it for subsequent transfer to the second and third layers. This preliminary staging allows the single image processing circuit to maintain higher effective processing speed by reducing idle wait time between processing different image pickup units.

Inventive Principle:
Principle #10Preliminary action

4Use of energy by moving object

If buffer size is reduced, then power consumption decreases, but data transfer efficiency may be compromised

Engineering Contradiction:
Improvepower consumptionVSAvoiddata transfer efficiency
Core Design Contradiction:
Use of energy by moving objectVSProductivity

Solution Approach 1:

The multilayer buffer structure segments the buffering function across three distinct layers, each with optimized capacity. This segmentation allows the total buffer memory to be distributed efficiently, reducing the peak memory access requirements and power consumption while maintaining adequate capacity for data transfer operations. Each layer can be independently managed for optimal transfer efficiency.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

By adding the layer dimension, the patent creates a three-dimensional buffer architecture (layer, row, column) that allows data to be organized and transferred more efficiently. This dimensional organization enables parallel transfer operations across layers, maintaining high data transfer efficiency even with reduced total buffer capacity compared to a single large buffer.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Data Source

PatentUS11616908B2Image processing apparatus and image processing method
Publication Date: 2023.03.28 KK TOSHIBA
  • US11616908B2 patent drawing
  • US11616908B2 patent drawing
  • US11616908B2 patent drawing

AI summary

An image processing apparatus according to an embodiment includes a preprocessing circuit and a distortion correction circuit configured to process, in a time division manner, a plurality of images generated by a plurality of image pickup units and an output buffer circuit configured to buffer the processed plurality of images in a unit of a block and add an identification tag including an address and an identification ID to the block.