Perceptual Image Hashing With Texture and Invariant Vector Distance for Copy Detection

Efficient Hashing Method Using 2D-2D PCA for Image Copy Detection

“…Main techniques [18] Sensitive SIFT + DWT [19] Sensitive SVD + CVA [20] Sensitive LBP + DCT + DWT [21] Sensitive Canny edge + DCT [25] Sensitive QLRBP + mean pooling [27] Within 5°RT + DWT + DFT [28] Within 5°RT + DCT [29] Within 5°ZMs + salient map [30] Within 5°SVD + LBP [31] Within 10°LRSD + CS [33] Within 5°Canny edge + DWT [35] Within 10°Hu moments + DCT [36] Within 10°GLCM + DCT [39] Within 5°QSVD + PCS [40] Within 5°2D PCA + salient map [43] arbitrarily angle Visual model + SIFT [46] arbitrarily angle CVA + DCT [47] arbitrarily angle RP + entropy [48] arbitrarily angle MDS + LPT [51] arbitrarily angle DCT + DNS [52] arbitrarily angle FMT + fractal coding [53] arbitrarily angle KAZE feature + SVD…”

“…Liu and Huang's method is robust to rotation of 10°. In another work, a global feature called GLCM (Gray-Level Co-occurrence Matrix) and local features with DCT coefficients are both used to come up with a novel hashing method in [36]. This hashing method is also robust to rotation of 10°and can be applied in copy detection.…”

“…To test the advantage of the proposed hashing method, some famous hashing methods are compared, including RW hashing method [17], QSVD hashing method [39], 2D PCA hashing method [40], MDS hashing method [48], GLCM-DCT hashing method [36] and RE hashing method [47]. The compared methods have been recently published in prestigious conferences or journals, and some of them are also dimension reduction-based methods, such as QSVD hashing method [39], 2D PCA hashing method [40] and MDS hashing method [48].…”

“…Similarly, every selected image is processed with the same manipulations to produce the actual query image. The wellknown tool called P-R curve [36] is selected to theoretically analyze copy detection performance. In the P-R curve, the y-axis is the Precision and the x-axis is the Recall.…”

A novel hashing scheme via image feature map and 2D PCA

“…[23] D W T + CVA 2014 [27] S V D + CVA 2017 [30] Canny operator + DCT 2019 [31] R P + visual salient region 2019 [33] F M T + fractal coding 2020 [34] G T E + DNS 2020 [35] V i s u a l m o d e l + Zernike moments + DCT 2020 [36] S V D + KAZE feature 2021 [37] G L C M + DCT 2021 selected robust features by combining the Canny edge operator and DCT coefficients to calculate a hash. This hashing scheme has a good application on industrial surveillance images, but it is only robust to a small number of digital operations.…”