[train hyper-parameters: Namespace(batch_size=8, epochs=10, lr=0.005, lrf=0.01)]
[epoch: 1]
global correct: 0.9499
precision: ['0.9506', '0.1667', '0.1084', '0.1356', '0.0559', '0.0364']
recall: ['0.9997', '0.0001', '0.0002', '0.0082', '0.0002', '0.0005']
IoU: ['0.9503', '0.0001', '0.0002', '0.0078', '0.0002', '0.0005']
mean IoU: 0.1598
[epoch: 2]
global correct: 0.9495
precision: ['0.9579', '0.2632', '0.2353', '0.2039', '0.1255', '0.0317']
recall: ['0.9976', '0.0001', '0.0136', '0.0987', '0.0499', '0.0024']
IoU: ['0.9556', '0.0001', '0.0130', '0.0712', '0.0370', '0.0023']
mean IoU: 0.1799
[epoch: 3]
global correct: 0.9515
precision: ['0.9698', '0.0000', '0.1934', '0.2150', '0.1779', '0.0639']
recall: ['0.9966', '0.0000', '0.0178', '0.2456', '0.1826', '0.0018']
IoU: ['0.9666', '0.0000', '0.0166', '0.1295', '0.0990', '0.0017']
mean IoU: 0.2022
[epoch: 4]
global correct: 0.9520
precision: ['0.9654', '0.0000', '0.1926', '0.2280', '0.1916', '0.1779']
recall: ['0.9980', '0.0000', '0.0033', '0.2057', '0.1512', '0.0011']
IoU: ['0.9636', '0.0000', '0.0033', '0.1212', '0.0923', '0.0011']
mean IoU: 0.1969
[epoch: 5]
global correct: 0.9530
precision: ['0.9736', '0.0000', '0.1823', '0.2173', '0.2034', '0.2699']
recall: ['0.9971', '0.0000', '0.0078', '0.3020', '0.2338', '0.0037']
IoU: ['0.9708', '0.0000', '0.0075', '0.1446', '0.1221', '0.0037']
mean IoU: 0.2081
[epoch: 6]
global correct: 0.9531
precision: ['0.9722', '0.0000', '0.1941', '0.2291', '0.1999', '0.2854']
recall: ['0.9973', '0.0000', '0.0062', '0.3057', '0.2084', '0.0084']
IoU: ['0.9697', '0.0000', '0.0060', '0.1507', '0.1136', '0.0083']
mean IoU: 0.2081
[epoch: 7]
global correct: 0.9531
precision: ['0.9758', '0.0000', '0.2015', '0.2220', '0.1975', '0.3028']
recall: ['0.9966', '0.0000', '0.0086', '0.3185', '0.2660', '0.0114']
IoU: ['0.9725', '0.0000', '0.0083', '0.1505', '0.1278', '0.0111']
mean IoU: 0.2117
[epoch: 8]
global correct: 0.9532
precision: ['0.9738', '0.0000', '0.2022', '0.2238', '0.2042', '0.3003']
recall: ['0.9971', '0.0000', '0.0090', '0.2894', '0.2516', '0.0146']
IoU: ['0.9710', '0.0000', '0.0087', '0.1444', '0.1270', '0.0141']
mean IoU: 0.2109
[epoch: 9]
global correct: 0.9532
precision: ['0.9757', '0.0000', '0.1961', '0.2201', '0.2017', '0.2957']
recall: ['0.9967', '0.0000', '0.0086', '0.3082', '0.2696', '0.0165']
IoU: ['0.9725', '0.0000', '0.0083', '0.1473', '0.1304', '0.0159']
mean IoU: 0.2124
[epoch: 10]
global correct: 0.9534
precision: ['0.9732', '0.0000', '0.2090', '0.2303', '0.2084', '0.2833']
recall: ['0.9974', '0.0000', '0.0073', '0.2876', '0.2480', '0.0170']
IoU: ['0.9707', '0.0000', '0.0071', '0.1466', '0.1277', '0.0163']
mean IoU: 0.2114
没有合适的资源?快使用搜索试试~ 我知道了~
温馨提示
基于Swin-Unet自适应多尺度训练、多类别分割、迁移学习:人体脊椎分割(6分割)【包含数据、代码、训练好的结果等等】 项目介绍:总大小215MB 网络仅仅训练了10个epochs,全局像素点的准确度达到0.95,miou为0.21。训练epoch加大的话,性能还会更加优越! 代码介绍: 【训练】train 脚本会自动训练,代码会自动将数据随机缩放为设定尺寸的0.8-1.2倍之间,实现多尺度训练。为了实现多分割项目,utils中的compute_gray函数会将mask灰度值保存在txt文本,并且自动为网络定义输出的channel 【介绍】学习率采用cos衰减,训练集和测试集的损失和iou曲线可以在run_results文件内查看,图像由matplotlib库绘制。除此外,还保存了训练日志,最好权重等,在训练日志可以看到每个类别的iou、recall、precision以及全局像素点的准确率等等 【推理】把待推理图像放在inference目录下,直接运行predict脚本即可,无需设定参数 具体参考README文件,小白均可使用
资源推荐
资源详情
资源评论
收起资源包目录
基于Swin-Unet自适应多尺度训练、多类别分割、迁移学习:人体脊椎分割(6分割)【包含数据、代码、训练好的结果等等】 (1375个子文件)
.gitignore 50B
u-net-me.iml 335B
T1_0249_S8_png_jpg.rf.2e870ffc2efc4fc296a5c84e1cfdb9c4.jpg 51KB
T1_0183_S8_png_jpg.rf.2ef42fc0650a305b5fb21b93fb51ed94.jpg 49KB
T1_0266_S8_png_jpg.rf.3ac8db81766634d3ba3cced95480072f.jpg 46KB
T1_0011_S8_png_jpg.rf.dc7ed1d42dd8e7df93d57271267f840b.jpg 45KB
T1_0162_S8_png_jpg.rf.3b1ed4e422deb6539e314bf8b61eceb7.jpg 45KB
T1_0048_S8_png_jpg.rf.69db460df8207fc85c17e9bbead41ef5.jpg 44KB
T1_0276_S8_png_jpg.rf.12ab592b9fafb50a9bdf52af51a425ae.jpg 44KB
T1_0223_S7_png_jpg.rf.de6b702252a41fa7e296740b3d196ed9.jpg 44KB
T1_0035_S8_png_jpg.rf.b96c0feef90f8844961e34602e7ae4a5.jpg 44KB
T1_0104_S7_png_jpg.rf.d7fb5987b2338a52b589a2cc207ae759.jpg 44KB
T1_0038_S8_png_jpg.rf.648928a33cfb1e1b69f51c98c7420ce2.jpg 43KB
T1_0217_S8_png_jpg.rf.c7d3dcc68f89e3fbc5b1a588fa0a568a.jpg 43KB
T1_0129_S8_png_jpg.rf.f9eeddf2f4b04e7f4a97f6d7f15ba146.jpg 43KB
T1_0358_S8_png_jpg.rf.2f31651faf1816bc9feb5a0c147a0e03.jpg 43KB
T1_0144_S8_png_jpg.rf.b7675ecb2a1204db6f0048c670cb9a49.jpg 42KB
T1_0181_S8_png_jpg.rf.4108772b042702ce0407079a0aa5505e.jpg 42KB
T1_0570_S7_png_jpg.rf.e35c59819e54ed028ad20aa577208f1e.jpg 42KB
T1_0036_S8_png_jpg.rf.db3ba0a60d0fac10ede40804d4faaf9f.jpg 42KB
T1_0041_S8_png_jpg.rf.c6e7658cd9b6639b42734d5d4e0ab347.jpg 42KB
T1_0130_S8_png_jpg.rf.697ff81bab7d95431892922aa046345c.jpg 42KB
T1_0139_S8_png_jpg.rf.38e10c944e56288d78a1f7f6e3222ab5.jpg 42KB
T1_0560_S8_png_jpg.rf.d5e77bc4da57a569143d43f45a7dd3e5.jpg 41KB
T1_0568_S8_png_jpg.rf.810122305c92a9714510ec18f57dd8cf.jpg 41KB
T1_0090_S8_png_jpg.rf.4c9cd4c0c4c57ed7cceb982041652f85.jpg 41KB
T1_0493_S8_png_jpg.rf.9740fc54a3e2db89d403437c3ba4eca9.jpg 41KB
T1_0119_S8_png_jpg.rf.a648b62f763ff7489f419786c70188d9.jpg 41KB
T1_0188_S8_png_jpg.rf.1736189242fc712614a4af5a4c5b33c9.jpg 41KB
T1_0272_S8_png_jpg.rf.8a770295b08a1c745bd7c2ec5cc0650c.jpg 41KB
T1_0573_S8_png_jpg.rf.f10e65c31fad02096c66244dfabc1e23.jpg 41KB
T1_0093_S8_png_jpg.rf.a653820ab0e249a5f807a8b65f074cb9.jpg 41KB
T1_0016_S8_png_jpg.rf.baa70d96e4773637c469c32a01bca24c.jpg 41KB
T1_0103_S8_png_jpg.rf.71c79727e4663766a5181db4cf13f6dd.jpg 40KB
T1_0166_S8_png_jpg.rf.9ce710a8214a0bea447905b108cac085.jpg 40KB
T1_0283_S8_png_jpg.rf.88326f5bcbcd5733bc529d50009c0b39.jpg 40KB
T1_0184_S8_png_jpg.rf.e767d134a253d48fd3e548d31ef44f66.jpg 40KB
T1_0204_S8_png_jpg.rf.b2aa766530ba3209da929212751f329f.jpg 40KB
T1_0046_S8_png_jpg.rf.ca3b7d6294ec36e05d2e60401223ed24.jpg 40KB
T1_0173_S7_png_jpg.rf.6c2132b13f033679893e271d3989d344.jpg 40KB
T1_0005_S8_png_jpg.rf.269f58836dd0e6e93c00234680511fd7.jpg 40KB
T1_0057_S8_png_jpg.rf.ca9ecf773b68f84bdebfdcf710732c13.jpg 40KB
T1_0042_S15_png_jpg.rf.4bd398ac9280663ecfede238a7a4a600.jpg 40KB
T1_0109_S8_png_jpg.rf.5158078370dccdc4ecf7e6b1d283a77d.jpg 40KB
T1_0575_S8_png_jpg.rf.2d424b401b118bc255658f07f33a3d1b.jpg 40KB
T1_0065_S15_png_jpg.rf.3a6424a4c11944f43dd8970d1f203eac.jpg 39KB
T1_0173_S7_png_jpg.rf.7124503c7b0edf5ebc332aa5ec64f72a.jpg 39KB
T1_0010_S8_png_jpg.rf.5c942f19f7eb1bc132bd59dc1e992c63.jpg 39KB
T1_0284_S8_png_jpg.rf.542a1b6d9b613516865725908dbad92d.jpg 39KB
T1_0051_S8_png_jpg.rf.a59ddf7e3a08ccbc7de055c1ff947d4f.jpg 39KB
T1_0154_S8_png_jpg.rf.37eccadab42ab7ff1afba4acb0eff736.jpg 39KB
T1_0564_S8_png_jpg.rf.f961a0a2e5f61b6b74dc0a991051b11e.jpg 39KB
T1_0254_S8_png_jpg.rf.6802242cb8b24f725605540ba49dc208.jpg 39KB
T1_0112_S8_png_jpg.rf.e56fd2063dd52619c027c57ac49e83b2.jpg 38KB
T1_0176_S8_png_jpg.rf.c6ae496f83511fcf5b69212f665cd245.jpg 38KB
T1_0013_S8_png_jpg.rf.3cff32f39603b0ec024ae2e30f6396e9.jpg 38KB
T1_0219_S8_png_jpg.rf.131dc8bd3584891130d68b7cd5b139d0.jpg 38KB
T1_0105_S8_png_jpg.rf.33f86b753d01271f9200656291b064d4.jpg 38KB
T1_0153_S8_png_jpg.rf.d082ddb6629cf8662293c49ec6256265.jpg 38KB
T1_0202_S8_png_jpg.rf.4a695aa8ac75da3256e12e6788af7eaa.jpg 38KB
T1_0099_S8_png_jpg.rf.1c7f577a93923783bd7fac42a935c719.jpg 38KB
T1_0115_S8_png_jpg.rf.d604eb73e95ca7a6852a9606c57a6509.jpg 38KB
T1_0138_S6_png_jpg.rf.896f4f269a0448531a20316d6bbb19aa.jpg 38KB
T1_0208_S8_png_jpg.rf.d3d15bc83df4e142aaa646a9bdf15e08.jpg 38KB
T1_0181_S8_png_jpg.rf.fa39a880df4cf9a977e272e905ba6315.jpg 38KB
T1_0281_S8_png_jpg.rf.d3aceceec55360902b225bad1fe7bea5.jpg 38KB
T1_0100_S8_png_jpg.rf.ccd31a1e27c6ac4a2a1ac6c70da168d1.jpg 37KB
T1_0074_S7_png_jpg.rf.5980839288e746cebf45fc946d8f9b8a.jpg 37KB
T1_0278_S8_png_jpg.rf.1e1801416aa778beeedb3fd8592f9d57.jpg 37KB
T1_0255_S8_png_jpg.rf.9e197c299e7843168298dbe13f11b5a2.jpg 37KB
T1_0559_S8_png_jpg.rf.587263b40688d5e1f2934c783681b6a9.jpg 37KB
T1_0027_S14_png_jpg.rf.573b80941f30501b0e759a2386539671.jpg 37KB
T1_0561_S7_png_jpg.rf.fc5f4f2b0ac713a66399d2eb0876b0ab.jpg 37KB
T1_0094_S7_png_jpg.rf.8599e436feb6e6d7f6137da6527f7c2c.jpg 37KB
T1_0099_S8_png_jpg.rf.fe63cb16cc2e73e8525332aa5d20d2f3.jpg 37KB
T1_0286_S8_png_jpg.rf.6671735f5b3afce4df466cf2ec7814c6.jpg 37KB
T1_0047_S8_png_jpg.rf.e8cf23518b7aa2a4b202e891fa109f8c.jpg 37KB
T1_0062_S8_png_jpg.rf.04d4b0301dab1ccf479a02dc7832b47e.jpg 37KB
T1_0207_S8_png_jpg.rf.48f3cbeef665d1cb699e26bcdec978cf.jpg 37KB
T1_0219_S8_png_jpg.rf.433c2a52b7991f45a8f87a5fadc90a8f.jpg 37KB
T1_0286_S8_png_jpg.rf.a1226911376b4a8a00903cc109fcffc3.jpg 37KB
T1_0121_S8_png_jpg.rf.ba778c66320c959689c44b9ccfee9441.jpg 37KB
T1_0292_S8_png_jpg.rf.5235b33034a951f88fa98566881efe8c.jpg 37KB
T1_0241_S8_png_jpg.rf.5d1683a0cc84290f6a383327fde93227.jpg 37KB
T1_0244_S7_png_jpg.rf.f115242b23cd84e0205a97732863e96a.jpg 37KB
T1_0194_S8_png_jpg.rf.57bc10601c7dceb8ad7bfc34abec23e2.jpg 37KB
T1_0244_S7_png_jpg.rf.e3f60734c592c5bb99cb51fa2c55512b.jpg 37KB
T1_0200_S7_png_jpg.rf.6a628da186ef0ab37a5655ca893b347e.jpg 36KB
T1_0081_S8_png_jpg.rf.49b89f6b516a482e22eaadb62c4f5fa4.jpg 36KB
T1_0268_S8_png_jpg.rf.74ab8847922d8e97a62bb567988f0c52.jpg 36KB
T1_0047_S8_png_jpg.rf.406cdf6954c9f7b5c485f33301eb0b5a.jpg 36KB
T1_0173_S7_png_jpg.rf.7ab85d804a2f42676e523c7ebbbb79c3.jpg 36KB
T1_0219_S8_png_jpg.rf.d7d3387926c7cebd5a98e721947048ed.jpg 36KB
T1_0176_S8_png_jpg.rf.6b2480ca8bde5e3213a59d2eff69d795.jpg 36KB
T1_0131_S7_png_jpg.rf.a8e4d046faf68c5b5f8c363042d1704e.jpg 36KB
T1_0232_S8_png_jpg.rf.f853cd79109bbd5b6aedbed9e25cee44.jpg 36KB
T1_0286_S8_png_jpg.rf.3883bb48aa2555f3a3a4c5253e4fde46.jpg 36KB
T1_0097_S8_png_jpg.rf.af3a3eb132c4f9c1a18de300db3ab0b3.jpg 36KB
T1_0265_S8_png_jpg.rf.c95a13c1ba74f8d3e5a6cbca13f41d52.jpg 36KB
T1_0245_S8_png_jpg.rf.50109c77aff076b361f9e29d1606221e.jpg 36KB
共 1375 条
- 1
- 2
- 3
- 4
- 5
- 6
- 14
资源评论
听风吹等浪起
- 粉丝: 1w+
- 资源: 1253
上传资源 快速赚钱
- 我的内容管理 展开
- 我的资源 快来上传第一个资源
- 我的收益 登录查看自己的收益
- 我的积分 登录查看自己的积分
- 我的C币 登录后查看C币余额
- 我的收藏
- 我的下载
- 下载帮助
安全验证
文档复制为VIP权益,开通VIP直接复制
信息提交成功