SpikeSegNet-a deep learning approach utilizing encoder-decoder network with hourglass for spike segmentation and counting in wheat plant from visual imaging

Table 3 Details of each decoder block and corresponding transpose convolutional layer

Decoder block #	Input to transpose convolution	Output of transpose convolution	Input to decoder block	Convolution filter size	Number of convolution filter	Output of decoder block
Block 1	32 * 32 * 128	64 * 64 * 128	64 * 64 * 128	3 * 3	128	64 * 64 * 128
Block 2	64 * 64 * 128	128 * 128 * 64	128 * 128 * 64	3 * 3	64	128 * 128 * 64
Block 3	128 * 128 * 64	256 * 256 * 64	256 * 256 * 64	3 * 3	16	256 * 256 * 16

ISSN: 1746-4811