Frame Interpolation Based on Visual Correspondence and Coherency Sensitive Hashing

Table 1

Comparison of average PSNR, RMSE, and MSSIM values.

Method	Evaluation	Walk	Jump-in-place	Silent	Space
TSS	PSNR	33.6385	37.1343	31.8166	34.1798
	RMSE	15.8233	11.4544	21.0495	15.3359
	MSSIM	0.9351	0.9402	0.9065	0.9647
ARPS	PSNR	34.2260	37.3584	32.0574	33.1645
	RMSE	15.8436	11.08734	20.4844	16.7526
	MSSIM	0.9393	0.9405	0.9125	0.9616
H&S	PSNR	27.7678	34.7418	29.3596	25.3645
	RMSE	31.5615	15.5741	29.3785	40.4711
	MSSIM	0.8989	0.9296	0.8757	0.8698
Classic+NL-Full	PSNR	28.3820	35.9997	29.9458	26.3099
	RMSE	30.4592	14.1522	27.1623	36.5059
	MSSIM	0.9210	0.9456	0.8953	0.8761
CSH	PSNR	29.2065	32.4474	32.9375	22.0820
	RMSE	26.4691	21.8114	18.8703	53.6209
	MSSIM	0.9325	0.9351	0.9275	0.9141
Proposed algorithm	PSNR	35.4886	38.4762	33.4151	34.8564
	RMSE	12.7313	9.7977	17.9540	14.3546
	MSSIM	0.9472	0.9518	0.9323	0.9717