Route background semantic X2 open-source image segmentation model - Free deployment to assist road background semantic segmentation

Route Background Semantic X2

Developed by Logiroad

This model is a fine-tuned image segmentation model based on Logiroad/route_background_semantic_x2 on the Logiroad/route_background_semantic dataset, primarily used for road background semantic segmentation tasks.

Image Segmentation

Transformers

#Road Defect Segmentation #Low-Light Optimization #Multi-Class Recognition

Downloads 78

Release Time : 4/4/2025

Model Overview

This is a vision model for road background semantic segmentation, capable of identifying and segmenting various road features in images, such as cuts, weather reflections, ice, or water seepage.

Model Features

Multi-Class Segmentation

Capable of identifying and segmenting multiple road features, including cuts, weather reflections, ice, or water seepage.

High-Precision Segmentation

Demonstrates high accuracy and intersection-over-union (IoU) across multiple segmentation categories.

Optimized Training

Trained using the AdamW optimizer and cosine learning rate scheduler to optimize model performance.

Model Capabilities

Image Segmentation

Road Feature Recognition

Semantic Segmentation

Use Cases

Road Maintenance

Road Damage Detection

Identifies road damages such as cuts, ice, or water seepage.

Achieves an accuracy of 0.2100 for the cuts category and 0.3102 for the ice or water seepage category.

Weather Reflection Analysis

Detects weather reflections on roads to help assess road safety.

Achieves an accuracy of 0.1023 for the weather reflection category.

Infrastructure Management

Road Repair Assessment

Identifies road areas requiring repairs to assist in maintenance planning.

Achieves an accuracy of 0.3946 for other repair categories.

🚀 route_background_semantic_x2

This model is a fine - tuned version of Logiroad/route_background_semantic_x2 on the Logiroad/route_background_semantic dataset, aiming at image segmentation tasks.

🚀 Quick Start

This model is a fine - tuned version of Logiroad/route_background_semantic_x2 on the Logiroad/route_background_semantic dataset. It achieves the following results on the evaluation set:

Loss: 0.4592
Mean Iou: 0.2260
Mean Accuracy: 0.2982
Overall Accuracy: 0.3342
Accuracy Unlabeled: nan
Accuracy Découpe: 0.2100
Accuracy Reflet météo: 0.1023
Accuracy Autre réparation: 0.3946
Accuracy Glaçage ou ressuage: 0.3102
Accuracy Emergence: 0.4740
Iou Unlabeled: 0.0
Iou Découpe: 0.1894
Iou Reflet météo: 0.1009
Iou Autre réparation: 0.3656
Iou Glaçage ou ressuage: 0.2399
Iou Emergence: 0.4600

📚 Documentation

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

🔧 Technical Details

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 6e - 05
train_batch_size: 6
eval_batch_size: 6
seed: 1337
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon = 1e - 08 and optimizer_args = No additional optimizer arguments
lr_scheduler_type: cosine
training_steps: 200000

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Unlabeled	Accuracy Découpe	Accuracy Reflet météo	Accuracy Autre réparation	Accuracy Glaçage ou ressuage	Accuracy Emergence	Iou Découpe	Iou Reflet météo	Iou Autre réparation	Iou Glaçage ou ressuage	Iou Emergence
0.1536	1.0	1618	0.3285	0.1946	0.2516	0.2839	nan	0.3153	0.0910	0.3010	0.1459	0.4049	0.2603	0.0871	0.2855	0.1365	0.3979
0.1476	2.0	3236	0.3270	0.2135	0.2824	0.2909	nan	0.2356	0.0396	0.3264	0.2896	0.5205	0.2038	0.0381	0.3074	0.2251	0.5063
0.133	3.0	4854	0.3192	0.2185	0.2969	0.2849	nan	0.2856	0.0318	0.2957	0.3583	0.5131	0.2490	0.0310	0.2801	0.2491	0.5021
0.1196	4.0	6472	0.3359	0.2116	0.2695	0.2911	nan	0.2549	0.1695	0.3208	0.1817	0.4206	0.2237	0.1600	0.3062	0.1659	0.4136
0.135	5.0	8090	0.3212	0.1941	0.2522	0.2972	nan	0.1861	0.2104	0.3464	0.2456	0.2725	0.1775	0.1884	0.3222	0.2075	0.2691
0.1131	6.0	9708	0.3212	0.2149	0.2833	0.2909	nan	0.1995	0.0422	0.3379	0.2808	0.5560	0.1875	0.0419	0.3168	0.2055	0.5378
0.1302	7.0	11326	0.3247	0.2326	0.3296	0.2937	nan	0.1116	0.2272	0.3412	0.4223	0.5458	0.1035	0.2134	0.3139	0.2360	0.5290
0.1277	8.0	12944	0.3144	0.2320	0.3045	0.2709	nan	0.2661	0.1839	0.2734	0.2839	0.5150	0.2167	0.1800	0.2618	0.2326	0.5009
0.1222	9.0	14562	0.3055	0.2187	0.2814	0.2922	nan	0.1104	0.1301	0.3606	0.2939	0.5121	0.1078	0.1270	0.3341	0.2447	0.4988
0.1258	10.0	16180	0.3037	0.2630	0.3598	0.3654	nan	0.3144	0.2259	0.3946	0.3186	0.5456	0.2612	0.1954	0.3715	0.2214	0.5285
0.1214	11.0	17798	0.3165	0.2267	0.2981	0.2951	nan	0.1908	0.1475	0.3383	0.2773	0.5366	0.1717	0.1424	0.3123	0.2133	0.5205
0.1495	12.0	19416	0.3099	0.2153	0.2804	0.2703	nan	0.2186	0.1025	0.2965	0.2727	0.5117	0.1936	0.0973	0.2800	0.2188	0.5024
0.1455	13.0	21034	0.3158	0.2191	0.2792	0.2912	nan	0.2227	0.1164	0.3343	0.1687	0.5537	0.2003	0.1105	0.3133	0.1529	0.5374
0.1207	14.0	22652	0.3487	0.2099	0.2877	0.2788	nan	0.1758	0.0801	0.3184	0.3497	0.5147	0.1602	0.0776	0.2989	0.2403	0.4826
0.1377	15.0	24270	0.3429	0.2285	0.2958	0.3017	nan	0.2715	0.0783	0.3323	0.2284	0.5685	0.2310	0.0755	0.3111	0.2018	0.5516
0.1017	16.0	25888	0.3044	0.2232	0.2861	0.3146	nan	0.2510	0.0860	0.3632	0.1736	0.5566	0.2181	0.0811	0.3415	0.1635	0.5349
0.1143	17.0	27506	0.3295	0.1975	0.2650	0.2489	nan	0.1555	0.1185	0.2812	0.3012	0.4687	0.1383	0.1116	0.2649	0.2162	0.4539
0.1178	18.0	29124	0.3098	0.2177	0.2892	0.3239	nan	0.2205	0.0896	0.3835	0.2163	0.5359	0.1929	0.0853	0.3525	0.1537	0.5220
0.1088	19.0	30742	0.3152	0.2153	0.2910	0.3126	nan	0.3170	0.0881	0.3294	0.3120	0.4085	0.2638	0.0835	0.3131	0.2343	0.3970
0.1179	20.0	32360	0.3254	0.2003	0.2557	0.2795	nan	0.1022	0.1233	0.3542	0.1883	0.5105	0.0989	0.1192	0.3229	0.1651	0.4960
0.1237	21.0	33978	0.3498	0.2061	0.2710	0.2720	nan	0.2235	0.0811	0.3013	0.2573	0.4918	0.1956	0.0801	0.2871	0.1908	0.4828
0.1053	22.0	35596	0.3051	0.2482	0.3256	0.3249	nan	0.2774	0.2052	0.3533	0.2449	0.5473	0.2481	0.1926	0.3347	0.1835	0.5300
0.1246	23.0	37214	0.3448	0.2283	0.2942	0.3091	nan	0.2029	0.1376	0.3601	0.2433	0.5272	0.1901	0.1290	0.3371	0.2034	0.5102
0.1351	24.0	38832	0.3383	0.1865	0.2490	0.3183	nan	0.1811	0.0292	0.3895	0.3113	0.3339	0.1647	0.0290	0.3641	0.2312	0.3300
0.1007	25.0	40450	0.3507	0.1966	0.2529	0.2824	nan	0.1047	0.0695	0.3562	0.2720	0.4619	0.0971	0.0685	0.3307	0.2286	0.4548
0.1173	26.0	42068	0.3579	0.2311	0.3147	0.3175	nan	0.3686	0.0762	0.3216	0.2766	0.5306	0.2827	0.0753	0.3083	0.2020	0.5185
0.1451	27.0	43686	0.3641	0.2351	0.3139	0.2743	nan	0.2167	0.2535	0.2887	0.2743	0.5362	0.1899	0.2141	0.2740	0.2171	0.5155
0.1412	28.0	45304	0.3442	0.2226	0.2927	0.3329	nan	0.2880	0.0836	0.3740	0.2592	0.4588	0.2455	0.0827	0.3459	0.2091	0.4524
0.1147	29.0	46922	0.3662	0.1812	0.2335	0.2758	nan	0.1350	0.0095	0.3448	0.2444	0.4336	0.1272	0.0095	0.3165	0.2057	0.4285
0.1171	30.0	48540	0.3362	0.2031	0.2592	0.2895	nan	0.2038	0.0681	0.3403	0.2106	0.4734	0.1799	0.0669	0.3197	0.1904	0.4618
0.092	31.0	50158	0.3502	0.2403	0.3151	0.3167	nan	0.1925	0.2658	0.3605	0.2906	0.4660	0.1711	0.2469	0.3321	0.2336	0.4580
0.1144	32.0	51776	0.3259	0.2194	0.2835	0.3300	nan	0.1992	0.0378	0.4015	0.2510	0.5278	0.1851	0.0374	0.3735	0.2041	0.5164
0.1063	33.0	53394	0.3323	0.2468	0.3224	0.3813	nan	0.1410	0.2256	0.4799	0.2847	0.4808	0.1277	0.2172	0.4326	0.2418	0.4612
0.0905	34.0	55012	0.3525	0.2342	0.3062	0.3177	nan	0.1441	0.2233	0.3825	0.2678	0.5135	0.1343	0.2136	0.3507	0.2125	0.4938
0.1112	35.0	56630	0.3824	0.1652	0.2235	0.2693	nan	0.0766	0.0035	0.3517	0.2814	0.4043	0.0745	0.0035	0.3252	0.1894	0.3983
0.1043	36.0	58248	0.3510	0.1874	0.2424	0.2676	nan	0.1363	0.1159	0.3228	0.2512	0.3860	0.1249	0.1086	0.2997	0.2101	0.3813
0.0946	37.0	59866	0.3340	0.2337	0.3170	0.3046	nan	0.2499	0.1323	0.3298	0.3350	0.5382	0.2200	0.1286	0.3105	0.2246	0.5188
0.1	38.0	61484	0.3437	0.2177	0.2853	0.3238	nan	0.1663	0.0574	0.4003	0.2527	0.5497	0.1514	0.0556	0.3648	0.1971	0.5372
0.0981	39.0	63102	0.3336	0.2382	0.3202	0.3120	nan	0.2613	0.2716	0.3331	0.2587	0.4764	0.2284	0.2479	0.3103	0.1892	0.4532
0.1124	40.0	64720	0.3413	0.2313	0.2967	0.2924	nan	0.1631	0.2792	0.3376	0.2198	0.4837	0.1491	0.2641	0.3166	0.1893	0.4687
0.0801	41.0	66338	0.3572	0.2438	0.3242	0.3384	nan	0.2931	0.1708	0.3720	0.2468	0.5383	0.2406	0.1637	0.3464	0.1905	0.5215
0.1144	42.0	67956	0.3530	0.2161	0.2753	0.3320	nan	0.2198	0.0688	0.3995	0.2190	0.4692	0.2014	0.0679	0.3698	0.1977	0.4598
0.0817	43.0	69574	0.3602	0.2335	0.3137	0.3095	nan	0.3497	0.1127	0.3126	0.2757	0.5180	0.2783	0.1074	0.2992	0.2122	0.5038
0.1063	44.0	71192	0.3908	0.2167	0.2789	0.2958	nan	0.2053	0.0701	0.3455	0.2354	0.5382	0.1853	0.0669	0.3270	0.2049	0.5159
0.0763	45.0	72810	0.3780	0.2017	0.2672	0.2916	nan	0.1444	0.0337	0.3584	0.2827	0.5170	0.1358	0.0336	0.3271	0.2123	0.5015
0.0836	46.0	74428	0.3590	0.2065	0.2705	0.3276	nan	0.1314	0.0617	0.4173	0.2690	0.4729	0.1221	0.0613	0.3797	0.2120	0.4637
0.1067	47.0	76046	0.3186	0.2449	0.3231	0.3483	nan	0.2393	0.2747	0.3961	0.2597	0.4459	0.2053	0.2541	0.3662	0.2085	0.4354
0.1018	48.0	77664	0.3630	0.1991	0.2561	0.3048	nan	0.1744	0.0343	0.3754	0.2208	0.4758	0.1558	0.0340	0.3474	0.1904	0.4670
0.1003	49.0	79282	0.3545	0.1945	0.2490	0.3086	nan	0.1174	0.1484	0.3910	0.2321	0.3562	0.1115	0.1369	0.3612	0.2184	0.3444

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご