Research on Identification of Wild Mushroom Species Based on Improved Xception Transfer Learning

Fig. 1. Sample data of wild mushroom images. (a) Amanita exitalis; (b) Amanita fuliginea; (c) Amanita neoovoidea; (d) Amanita parvipantherina; (e) Amanita rubrovolvata; (f) Entoloma quadratum; (g) Panaeolus sphinctrinus; (h) Psilocybe coprophila; (i) Gyromitra infula; (j) Lonomidotis frondosa

Download full size

Fig. 2. Effects of different data enhancement methods. (a) Origin image; (b) random rotation; (c) horizontal flip; (d) color dither; (e) Gaussian noise; (f) histogram equalization; (g) random cut

Download full size

Fig. 3. Structural diagram of Xception

Download full size

Fig. 4. Experimental flow chart of wild mushroom species identification model

Download full size

Fig. 5. Principle diagram of CBAM's realization

Download full size

Fig. 6. Comparison of three kinds of neural network structures. (a) Traditional neural network; (b) Dropout neural network; (c) Disout neural network

Download full size

Fig. 7. Comparison among model parameters for different training methods. (a) Accuracy; (b) average training time

Download full size

Table 1. Structures of A, B, C, and D components in Xception

View table

Table 1. Structures of A, B, C, and D components in Xception

A	B	C	D
[Input 299×299×3]	Sep Conv 128, 3×3	$\{\begin{array}{l} ReLU \\ Sep Conv 256, 3 \times 3 \end{array}\}$ ×3	$\{\begin{array}{l} ReLU \\ Sep Conv 728, 3 \times 3 \end{array}\}$ ×3
Conv 32,3×3,stride of 2×2ReLU	ReLU
Conv 32,3×3,stride of 2×2ReLU	Sep Conv 128,3×3	MaxPool 3×3,stride of 2×2	MaxPool 3×3,stride of 2×2
	MaxPool 3×3,stride of 2×2

Table 2. Structures of E, F, and G components in Xception
View table
Table 2. Structures of E, F, and G components in Xception
E(repeated 8 times) F G
$\{\begin{array}{l} ReLU \\ Sep Conv 728, 3 \times 3 \end{array}\}$ ×3 ReLU Sep Conv 1536,3×3
Sep Conv 728,3×3 ReLU
ReLU Sep Conv 2048,3×3
Sep Conv 1024,3×3 ReLU
MaxPool 3×3,stride of 2×2 Global average pool

Table 3. Comparison among different feature map disturbance forms
View table
Table 3. Comparison among different feature map disturbance forms
Experimental number Training structure Top 1 /% Top 5 /%
1# Origin model 94.91 99.30
2# Dropout 95.92 99. 33
3# Disout 96.32 99.61

Table 4. Comparisonamong different training methods

View table

Table 4. Comparisonamong different training methods

Experimental number	Training method	Top 1 /%	Top 5 /%	T_Average/s
1##	Random parameters	92.10	98.10	1688.37
2##	Freezing all parameters	87.26	98.34	599.38
3##	Freezing partial parameters	96.47	99.69	647.91
4##	Training all network layers	97.02	99.67	1645.39

Table 5. Comparison among different proportions
View table
Table 5. Comparison among different proportions
Experimental number Size of training set Size of validation set Top 1 /% Top 5 /%
1-1 5 5 94.23 98.91
2-1 6 4 95.15 98.99
3-1 7 3 95.67 99.36
4-1 8 2 96.32 99.61

Table 6. Comparison among different model experiments

View table

Table 6. Comparison among different model experiments

Model	Top 1 /%	Top 5 /%	Number of parameters /10⁶
Alex	86.24	97.94	61.10
ResNet50	89.39	98.64	25.64
ResNet101	91.64	98.69	44.71
ResNet152	92.36	99.32	60.42
InceptionV1	90.64	98.85	13.02
InceptionV3	93.16	99.09	23.93
InceptionResNetV2	95.41	99.28	55.87
Xception	95.58	99.43	22.99
Dis-Xception-CBAM	96.32	99.61	24.04

Tools

Get Citation

Copy Citation Text

Degang Chen, Zieguli Ai, Pengbo Yin, Yanuo Lu, Shunping Li. Research on Identification of Wild Mushroom Species Based on Improved Xception Transfer Learning[J]. Laser & Optoelectronics Progress, 2021, 58(8): 0810023

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites