Skip to main content

Table 5 Comparison of inference speed and GPU memory consumption for end-to-end nuclear segmentation and classification

From: Instance-level semantic segmentation of nuclei based on multimodal structure encoding

Model

input size

inference time (ms)

GPU memory (GB)

Batch size

Ours

270\(\times\)270

45.6

8.4

16

HoVer-Net [53]

270\(\times\)270

82.3

11.2

16

Mask2former [54]

270\(\times\)270

63.7

9.8

16

Triple U-net [55]

270\(\times\)270

58.9

10.5

16

  1. All measurements were conducted on a single NVIDIA RTX 4090 GPU. Inference time is averaged over 1000 runs. GPU memory consumption includes model parameters and intermediate features
Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us