Table 4. Retrieval accuracy of 1, 3, 5, 7, and 9 CBIR image retrievals with traditional KNN classification and handcrafted features for each 10-fold.

Fold	Retrieve 1 Image		Retrieve 3 Images		Retrieve 5 Images		Retrieve 7 Images		Retrieve 9 Images
	Chebychev	Cosine	Chebychev	Cosine	Chebychev	Cosine	Chebychev	Cosine	Chebychev	Cosine
1	65.22%	82.61%	68.84%	68.84%	64.35%	69.57%	62.42%	68.94%	60.87%	66.67%
2	66.67%	66.67%	60.00%	63.70%	55.11%	56.89%	52.38%	55.87%	52.17%	55.80%
3	67.39%	71.74%	58.70%	63.04%	56.09%	57.39%	53.73%	55.90%	50.97%	52.90%
4	57.78%	55.56%	59.26%	53.33%	57.78%	58.22%	54.60%	54.60%	58.02%	57.78%
5	66.67%	68.89%	69.63%	71.85%	66.22%	67.11%	61.90%	61.27%	61.70%	62.28%
6	55.26%	60.53%	71.93%	62.28%	66.32%	61.05%	64.66%	60.15%	57.31%	59.65%
7	63.16%	63.16%	59.65%	57.89%	58.95%	57.37%	59.40%	56.39%	59.40%	56.39%
8	65.79%	60.53%	63.16%	61.40%	65.26%	62.11%	64.29%	58.65%	62.28%	58.77%
9	65.79%	63.16%	65.79%	63.16%	66.84%	62.11%	64.29%	61.65%	62.87%	61.99%
10	63.04%	63.04%	65.22%	62.32%	62.61%	60.00%	63.04%	60.87%	60.53%	59.65%
Average	63.68%	65.59%	64.22%	62.78%	61.95%	61.18%	60.07%	59.43%	58.61%	59.19%
Standard Deviation	3.68%	6.81%	4.35%	4.63%	4.11%	3.87%	4.31%	3.81%	3.72%	3.50%