Table 3.

Convolutional layer channel counts of the segmentation models before and after optimization

Convolutional layer	UConvNeXt	AAUConvNeXt	Reduction ratio	Convolutional layer	UConvNeXt	AAUConvNeXt	Reduction ratio
DSCL 01	32	37	−15.63%	DSCL 23	1,024	1,469	−43.46%
DSCL 02	128	141	−10.16%	DSCL 24	256	166	35.16%
DSCL 03	32	44	−37.50%	DSCL 25	1,024	1,382	−34.96%
DSCL 04	64	41	35.94%	DSCL 26	256	231	9.77%
DSCL 05	256	267	−4.30%	DSCL 27	256	176	31.25%
DSCL 06	64	46	28.13%	DSCL 28	1,024	703	31.35%
DSCL 07	256	301	−17.58%	DSCL 29	256	214	16.41%
DSCL 08	64	96	−50.00%	DSCL 30	1,024	540	47.27%
DSCL 09	128	118	7.81%	DSCL 31	256	303	−18.36%
DSCL 10	512	640	−25.00%	USCL 01	128	172	−34.38%
DSCL 11	128	158	−23.44%	USCL 02	512	418	18.36%
DSCL 12	512	697	−36.13%	USCL 03	128	169	−32.03%
DSCL 13	128	134	−4.69%	USCL 04	64	45	29.69%
DSCL 14	256	383	−49.61%	USCL 05	256	351	−37.11%
DSCL 15	1,024	853	16.70%	USCL 06	64	94	−46.88%
DSCL 16	256	337	−31.64%	USCL 07	32	36	−12.50%
DSCL 17	1,024	656	35.94%	USCL 08	128	101	21.09%
DSCL 18	256	361	−41.02%	USCL 09	32	28	12.50%
DSCL 19	1,024	674	34.18%	USCL 10	32	24	25.00%
DSCL 20	256	372	−45.31%	USCL 11	128	135	−5.47%
DSCL 21	1,024	681	33.50%	USCL 12	32	24	25.00%
DSCL 22	256	156	39.06%