This Java Project uses DCT(Discrete Cosine Transform) to reduce the frequencies and Hamming Distance to compute similarity between two Images.
Reference: www.hackerfactor.com
- Reduce Size : The input image is converted to 32x32 px.
- Reduce Color : The image is reduced to grayscale.
- Compute DTC : Computer the DTC of 32x32px image. DTC.java available on https://github.com/AshishKaktan/Discrete-Cosine-Transform/
- Reduce DTC : Just Keep the top left 8x8 matrix for further computation, this matrix represent the lowest frequencies.
- Compute the average value: Compute the average value of the Matrix excluding the First Value.
- Assigning Bits: Now Depending on whether each of the 64 DCT values is above or below the average set it to 0 or 1.
- Construct the hash: The resultant set of 64 bits is the generated hash for the image.