System and method for feature sub-image detection and identification in a given image
The present invention provides a system and method for detection and identification of feature sub-image in a given image. The system ( 100 ) comprises single CPY core with Single Instruction Multiple Data (SIMD) instruction set but the instructions lack saturation logic ( 101 ), wherein the software modules are implemented. Removing the logic hardware used to implement saturation in the SIMD instructions helps in lowering the BoM cost and the inventive steps helps in achieving bit-exact results without saturation logic. The power consumed using CPU without saturation logic in SIMD instruction is lesser than using Digital Signal Processing Instruction Set Architecture with saturation logic, thus giving value additions to platform SoC designers and makers for using low BoM cost solution.
1 . A system ( 100 ) for detection and identification of feature sub-image in a given image, the system ( 100 ) comprising:
a. a single Central Processing Unit (CPU) core with an Arithmetic Logic Unit (ALU) configuration and a non-transitory memory that stores software code, and Single Instruction Multiple Data (SIMD) instructions for executing software code for detection and identification of feature sub-image in a given image ( 101 );
b. Principal component analysis algorithm and inferencing technique stored on the non-transitory memory to detect and identify feature sub-image in the given image;
c. saturation logic is turned off in the SIMD instructions of the software code; and
d. the Arithmetic Logic Unit (ALU) of CPU is designed to have SIMD instructions without saturation logic in a critical instruction, wherein the critical instruction include Multiply and Accumulate (MAC) and shift instructions, wherein a Bill of Material (BoM) unit cost of a platform SoC consisting of CPU cores, DSP cores, and a GPU core is reduced compared to using a Digital Signal Processor (DSP) core to implement software modules by reducing a quantitative gate reduction by only including the CPU core and the GPU core for the platform SoC.
2 . A system ( 100 ) for detection and identification of feature sub-image in a given image, the system ( 100 ) comprising:
a. a single Central Processing Unit (CPU) core with an Arithmetic Logic Unit (ALU) configuration and a non-transitory memory that stores software code, and Single Instruction Multiple Data (SIMD) instructions for executing software code for detection and identification of feature sub-image in a given image ( 101 );
b. Principal component analysis algorithm and inferencing technique stored on the non-transitory memory to detect and identify feature sub-image in the given image;
c. saturation logic is turned off in the SIMD instructions of the software code; and
d. the Arithmetic Logic Unit (ALU) of CPU is designed to have SIMD instructions without saturation logic in a critical instruction, wherein the critical instruction include Multiply and Accumulate (MAC) and shift instructions, wherein a Bill of Material (BoM) unit cost of a platform SoC consisting of CPU cores, DSP cores, and a GPU core is reduced compared to using a Digital Signal Processor (DSP) core to implement software modules by reducing a quantitative gate reduction by only including the CPU core and the GPU core for the platform SoC.