Intelligent Plant Disease Diagnosis with Explainable AI Methods and

Lightweight Model

Ishan Joshi, Naman Mardia and R. Vidhya

CTech, SRM Institute of Science and Technology, Chennai, Tamil Nadu, India

Keywords: Plant Disease Detection, Artificial Intelligence (AI), Deep Learning, Explainable AI (XAI), Sustainable

Agriculture.

Abstract: Agriculture is a most important contributor to a country economy and especially in India, as majority of rural

people, its only source of livelihood. Plant diseases are among the most significant challenges to agriculture,

which can be caused by pathogen, synthetic fertilizers, outdated farming practices, and environmental

conditions. The yield for crops can be greatly reduced by these diseases that lead to substantial coronavirus

economic impact. AI and Machine Learning techniques for outbreak detection have become widely used by

researchers to tackle this issue. This survey examines prevalent plant leaf diseases, explores traditional and

deep learning methods for disease detection, and reviews available datasets. It also addresses the use of

Explainable AI (XAI) applied to deep learning to enhance the transparency of the models, leading to

understandable decisions for the user. Drawing on this expertise, the survey provides insights for researchers,

practitioners, and other stakeholders, informative the creation of effective and transparent biosolutions to plant

diseases, resulting in sustainable agricultural systems.

1 INTRODUCTION

Even with the advancement of agricultural

technology, plant disease continues to be a pressing

challenge, leading to increasing annual crop losses

globally and posing a significant threat to food

security (Shirahatti et al.; Ko Ko Zaw et al.;

Owomugisha et al.). Conventional methods for

detecting plant diseases heavily depend on expert

knowledge, which may introduce errors, biases, and

inefficiencies, especially in large-scale agricultural

practices (Hasib et al.; Singh et al.). These limitations

often delay treatment, allowing diseases to spread

further and reduce yields. The growing field of

machine learning and computer vision offers a

solution through automated systems for disease

detection (Cap et al.; Amin et al.). However, many

existing models’ function as black-box systems,

offering little to no insight into their decision-making

processes (Baehrens et al.; Lundberg and Lee). This

lack of transparency can reduce trust and hinder

adoption among farmers and agronomists who require

clear and verifiable recommendations (Wei et al.;

Daglarli). Additionally, such models may perform

poorly in varying visual environments (Rajeena et al.;

Sapura et al.). This underscores the need for intelligent

plant disease diagnosis systems that not only achieve

high accuracy but also incorporate explainability

through Explainable AI (XAI) methods (Arvind et al.;

Tabbakh and Barpanda). These systems must provide

interpretable outputs to empower users in making

informed decisions about disease management (Wei et

al.; Baehrens et al.). Furthermore, the solutions should

be robust, scalable, and adaptable to different

agricultural contexts to ensure widespread usability

(Cap et al.; Amin et al.). Therefore, this survey paper

presents a comprehensive overview of common leaf-

based plant diseases, available datasets, and state-of-

the-art detection techniques (Shirahatti et al.; Singh et

al.; Ko Ko Zaw et al.). It also highlights the application

of XAI techniques, especially in CNN and

Transformer models, to enhance interpretability and

model transparency in disease classification tasks

(Tabbakh and Barpanda; Arvind et al.; Lundberg and

Lee; Wei et al.). The paper emphasizes the importance

of XAI in this domain and outlines potential directions

for future research (Daglarli; Baehrens et al.; Amin et

al.).

Joshi, I., Mardia, N. and Vidhya, R.

Intelligent Plant Disease Diagnosis with Explainable AI Methods and Lightweight Model.

DOI: 10.5220/0013904700004919

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies (ICRDICCT‘25 2025) - Volume 3, pages

739-744

ISBN: 978-989-758-777-1

739

2 OVERVIEWS ON LEAF BASED

PLANT DISEASE DETECTION

All living organisms, including plants, animals, and

humans, are vulnerable to diseases. Researchers and

professionals in agricultural science and management

are actively searching for advanced solutions to

mitigate plant disease outbreaks, which can cause

significant damage to agricultural productivity

(Shirahatti et al.; Owomugisha et al.). To address this,

various scientific disciplines collaborate to control

the spread of plant leaf diseases and ensure a stable

food supply for the world’s growing population (Ko

Ko Zaw et al.; Sapura et al.; Rajeena et al.).

Plant diseases can manifest through various

symptoms that affect a plant’s structural components

such as leaves, stems, and roots, ultimately

influencing its growth and yield (Singh et al.; Cap et

al.). The occurrence of these diseases varies

seasonally, influenced by changes in weather

conditions and the presence of specific pathogens

(Amin et al.; Tabbakh and Barpanda). Recent

approaches using convolutional neural networks and

vision transformers show promising results in

improving detection rates (Arvind et al.; Wei et al.).

The integration of Explainable AI (XAI) further

enhances these models by offering transparent

insights into classification decisions, promoting trust

among end-users like farmers and agronomists

(Baehrens et al.; Lundberg and Lee; Daglarli).

Furthermore, research demonstrates that XAI models

such as SHAP, LIME, and attention-based techniques

are helping interpret deep learning decisions in

agriculture and beyond (Hasib et al.; Wei et al.). This

section thus surveys common leaf diseases, key

datasets, and notable contributions in the area of leaf-

based plant disease detection.

2.1 Common Leaf Diseases in Plants

Plant diseases predominantly affect the leaves, but

can also impact the roots, stems, and fruits.

Among these, leaf diseases are the most prevalent

and are typically managed using fungicides,

bactericides, or resistant plant varieties. Below are

some of the most common leaf diseases:

• Blight: One of the most destructive plant

diseases, Blight has historically caused

significant damage, such as during the 1840s

potato famine. This fungal disease spreads in

warm, humid conditions through wind-borne

spores.

• Scab: This fungal disease is host-specific

and can infect individual plants. It is

prevalent in apple trees, where it initially

causes olive green spots on the leaves, which

eventually turn yellow before the leaves fall

off.

• Powdery Mildew: Common in shaded

areas, Powdery Mildew is easily

recognizable by the white powdery coating

on the upper surface of the leaves. This

disease spreads in humid conditions with

low soil moisture.

• Mosaic Virus: The mosaic virus affects

plants at a molecular level, commonly

infecting tomatoes, tobacco, and other

horticultural plants. Infected leaves develop

yellowish and whitish stripes.

• Marssonina Blotch: Caused by the fungus

• Marsonina Caronaria, this disease occurs in

high rainfall areas. Infected leaves develop

circular dark green patches that can turn dark

brown in severe cases.

• Black Spot: Another fungal disease, Black

Spot, creates round black spots on the upper

surface of leaves. It thrives in prolonged wet

conditions or when leaves remain moist for

extended periods.

• Frogeye Spot: Caused by the fungus

Cercospora Sojina, Frogeye Spot manifests

as purple spots on leaves during early spring,

which later develop into brownish rings

resembling a frog's eye.

• Rust: This easily identifiable fungal disease

causes brownish rusty spots on leaves and is

commonly found on apples, roses, and

tomatoes, especially during wet weather in

early spring.

Plant leaf diseases are one of the growing

challenges in agricultural productivity due to a wide

range of pathogens such as fungus, bacteria, viruses,

etc. These pathogens have different lifecycles and

environmental triggers, making disease management

a potentially complex and multi-factors challenge.

Exploring the conditions specific to diseases such

as Blight and Rust for example to focus methods of

intervention. For example, these diseases can cause

significant changes in plant physiology, and impede

photosynthesis and nutrient uptake, resulting in

reduced growth and yield, or even death in some

cases if not properly managed.

The dataset includes various plant diseases, such

as Apple Scab (Figure 1), Rose Black Spot (Figure 2),

Powdery Mildew (Figure 3), Blight (Figure 4),

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

740

Mosaic Virus (Figure 5), Marssonina Blotch (Figure

6), Frogeye Spot (Figure 7), and Rust (Figure 8), each

of which presents unique visual characteristics aiding

in accurate classification.

Figure 1: Apple scab.

Figure 2: Rose black spot.

Figure 3: Powdery mildew.

Figure 4: Blight.

Figure 5: Mosaic virus.

Figure 6: Marssonina blotch.

Figure 7: Frogeye spot.

Figure 8: Rust.

3 OVERVIEWS ON

EXPLAINABLE AI (XAI)

AI (Artificial Intelligence) is the talk of the town, and

with good reason, for almost every subtype of

research is either moving towards AI or refactor their

already existing rule-based system to AI systems. But

many of today’s AI systems, such as those that

employ deep learning and machine learning, are very

opaque, meaning that users cannot tell what is

happening inside the system or what drives important

decisions. This lack of source transparency creates

mistrust and discourages users from adopting your

final product.

Certain AI researchers claim that explanations do

not need to be an area of focus in AI research as it is

either not necessary or too ambitious, whereas others

have stated that rather than hinder human

intelligence, explanations accompanying AI outputs

can facilitate it and can build trust in these systems.

In this way, closing this gap would enhance trust on

Intelligent Plant Disease Diagnosis with Explainable AI Methods and Lightweight Model

741

AI systems and unleash opportunities for AI led

products & services.

AI systems need to explain things to users in

critical areas such as law, medicine, agriculture,

finance and defence so that they can apply them

safely and confidently. Explanations provide a useful

layer of human-computer interaction, enabling users

to get more value from AI-based services. AI has

been advancing rapidly, made possible in no small

part by machine learning methods ranging from

Support Vector Machines (SVMs) and Random

Forests (RF) to probabilistic models to Deep Learning

(DL) neural networks, all of which work as blackbox

models. These models require little to no human input

in order to run, and can be employed right away in

multiple environments with little tailoring.

But according to the tradition we have been

trading off the performance of machine learning

models such as predictive accuracy with model

interpretability. For example, deep learning is often

highly accurate models, but not very explainable,

while decision trees are very explainable, but not very

accurate. A hypothetical graph (Figure 2) illustrates

this performance-explainability trade-off,

demonstrating that explainability typically decreases

while model performance increases. Figure 9 Shows

Comparative Analysis of Machine Learning

Algorithms: Explainability vs Learning Performance.

To address this challenge and make AI solutions

more transparent and trustworthy, a research domain

called Explainable Artificial Intelligence (XAI) has

emerged. XAI aims to enhance the interpretability of

AI systems, making them easier for users to

understand and trust. A. What is XAI?

Figure 9: Comparative analysis of machine learning.

That said, artificial intelligence (AI) has become

a hot new commodity, bringing about a considerable

change in numerous environments, from automatic

vehicles to medical diagnostics. Yet users who are not

technical often do not understand the systems on

which they depend, and therefore, trust in AI-

generated decisions cannot be taken for granted. In

industries as sensitive as defences, healthcare, and

safety, this issue is being cast in an even graver light,

given the extent of AI's integration into these

domains. As AI becomes more prevalent in

supporting or even replacing human supervisors in

these domains, it is imperative to do more than show

how the AI reached a given decision; it is a

prerequisite of any responsible AI system to ensure

the users can verify how the system works and how

the AI works.

Concerns like these are why Explainable

Artificial Intelligence (XAI), a sub field of machine

learning, was created. These XAI techniques are

finding applications for improving the transparency

and reliability of AI systems by exposing their inner

workings in a manner that users can appreciate and

also be confident in the model’s decisions. By

implementing ethical considerations, XAI minimizes

the chances of unintentional bias while boosting

confidence in the system's results. XAI is mainly

focused on providing humans a set of rules for XAI

decisions and making AI systems more user and

efficient friendly according to some similar

principles.

For example, consider a healthcare scenario where

a patient with breathing issues is placed on a ventilator.

A doctor monitors the patient's heart rate through an

AIenabled system, which displays fluctuating heart

rates on the screen.

The AI algorithm is designed to predict the

patient's heart rate for the next 15 seconds based on

previous and current data. However, this system, like

many "black-box" models, provides highly accurate

predictions without explaining the factors influencing

these heart rate variations. In this case, the doctor is

relying on an AI system that offers no insight into its

decision-making process, making it risky to trust such

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

742

a system without understanding the internal factors

driving the predictions.

This hypothetical example highlights the need for

explainable AI systems in high-stakes environments,

where users must be able to trust and understand the

decisions made by AI in order to use them effectively

and safely. By making AI more transparent, XAI can

bridge this gap.

4 MODEL ARCHITECTURE

In this work, we used the LAAMA (Lightweight

Attention-

Augmented Mobile-friendly Architecture),

alongside EfficientNetV2L, MobileNetV2, and

ResNet152V2 for comparison, to detect 38 diseases

across 14 plant species. LAAMA incorporates

lightweight attention modules to improve feature

extraction and focus on important regions in the

images while keeping the model suitable for

deployment on mobile and edge devices.

The LAAMA architecture is designed with depth

wise separable convolutions and attention

mechanisms that reduce computational complexity

and make it mobile-friendly.

We pertained LAAMA on the ImageNet dataset,

fine-tuned it for plant disease detection, and followed

the same steps for the other models to ensure a fair

comparison.

We used the Adam optimizer with a learning rate

of 0.0001, categorical cross-entropy as the loss

function, and a softmax activation function in the

output layer, which has 38 neurons due to the

multiclass classification nature of this task. All

models, including LAAMA, were trained for 50

epochs, using a dropout function to mitigate

overfitting.

Table 1: Performance Metrics Comparison of Deep Learning

Models for Classification Tasks.

Model

LAA

EfficientNe

tV2L

MobileN

etV2

ResNet1

52V2

Accur

acy

99.25

99.63% 98.86% 98.44%

Precis

ion

99.13

99.63% 98.68% 98.19%

Recall

98.94

99.63% 98.03% 97.53%

Score

99.03

99.63% 98.29% 97.82%

5 RESULT ANALYSIS

We have tested our models on quantitative

performance evaluation metrics: accuracy (1),

precision (2), recall (3), and f1 score (4) by their

predictions on our test set. Figure 10 Shows the

Training and Validation Accuracy.

𝐴𝑐𝑐𝑢𝑟𝑎𝑐𝑦 =

 



(1)

Here, TN = True negative, TP = True positive, FN =

False negative, FP = False positive.

𝑃𝑟𝑒𝑐𝑖𝑠𝑖𝑜𝑛 =





(2)

Here, TP = True positive, FP = False positive.

𝑅𝑒𝑐𝑎𝑙𝑙 =

∗∗

 

(3)

Here, TP = True positive, FN = False negative.

𝐹1𝑆𝑐𝑜𝑟𝑒 =

∗∗

 

(4)

Figure 10: Training and Validation Accuracy.

From Table 1, we observe that EfficientNetV2L

achieved the highest performance across all metrics,

but LAAMA provided competitive results while

maintaining a lightweight architecture suitable for

mobile deployment:

• Accuracy: LAAMA scored 99.25%, which is

0.38% lower than EfficientNetV2L but higher

than both MobileNetV2 and ResNet152V2.

• Precision: LAAMA achieved 99.13%

precision, 0.5% lower than EfficientNetV2L

but higher than MobileNetV2 and

ResNet152V2.

• Recall: LAAMA had a 98.94% recall, 0.69%

lower than EfficientNetV2L but still higher

Intelligent Plant Disease Diagnosis with Explainable AI Methods and Lightweight Model

743

than MobileNetV2 and ResNet152V2.

• F1 Score: LAAMA's F1 score was 99.03%,

just 0.60% lower than EfficientNetV2L and

higher than the other two models.

• Thus, while EfficientNetV2L outperformed in

raw accuracy and precision, LAAMA

provides a strong trade-off between

performance and mobile-friendliness, making

it highly suitable for applications requiring

real-time processing on edge devices.

REFERENCES

Adi Dwifana Sapura, Djarot Hindarto, Handri Santoso,

“Disease Classification on Rice Leaves using

DenseNet121, DenseNet169, DenseNet201”,

Sinkyone: Jurnal dan Penelitian Teknik Informatika

Volume 8, Issue 1, January 2023, DOI:

https://doi.org/10.33395/sinkron.v8i1.11906

Amer Tabbakh and Soubhagya Sankar Barpanda. “A Deep

Features Extraction Model Based on the Transfer

Learning Model and Vision Transformer “TLMViT”

for Plant Disease Classification”. In: IEEE Access 11

(2023). Conference Name: IEEE Access, pp. 45377–

45392. ISSN: 2169-3536

CS Arvind et al. “Deep Learning Based Plant Disease

Classification with Explainable AI and

Mitigation Recommendation”. In: 2021 IEEE

Symposium Series on Computational Intelligence

(SSCI). Dec. 2021, pp. 01– 08

D. Baehrens, T. Schroeter, S. Harmeling, M. Kawanabe, K.

Hansen, and K. R. Müller, “How to explain individual

classification decisions,” J. Mach. Learn. Res., vol. 11,

pp. 1803–1831, 2010

Daglarli, Evren, “Explainable Artificial Intelligence (xAI)

Approaches and Deep Meta-Learning Models for

Cyber-Physical Systems”, Artificial Intelligence

Paradigms for Smart Cyber-Physical Systems, edited

by Ashish Kumar Luhach and Atilla Elçi, IGI Global,

2021, pp. 42-67.

Fathimathul Rajeena, Aswathy S, Mohamed A. Moustafa

and Mona A. S. Ali, “Detecting Plant Disease in Corn

Leaf Using EfficientNet Architecture An Analytical

Approach”, Electronics 2023, https://doi.org/10.3390/

electronics12081938

Godliver Owomugisha, Friedrich Melchert, Ernest

Mwebaze, John A Quinn and Michael Biehl, “Machine

Learning for diagnosis of disease in plants using

spectral data”, Int'l Conf. Artificial Intelligence (2018).

Godliver Owomugisha, Friedrich Melchert, Ernest

Mwebaze, John A Quinn and Michael Biehl, “Machine

Learning for diagnosis of disease in plants using

spectral data”, Int'l Conf. Artificial Intelligence (2018).

HASSAN AMIN, ASHRAF DARWISH (Member, IEEE),

ABOUL ELLA HASSANIEN AND MONA

SOLIMAN. “End-to-End Deep Learning Model for

Corn Leaf Disease Classification”, IEEE Access, “,

Volume 10, 2022, Digital Object Identifier

10.1109/ACCESS.2022.3159678

J. Shirahatti, R. Patil, and P. Akulwar, “A survey paper on

plant disease identification using machine learning

approach,” in 2018 3rd International Conference on

Communication and Electronics Systems (ICCES).

IEEE, 2018, pp. 1171–1174.

K. M. Hasib, F. Rahman, R. Hasnat, and M. G. R. Alam,

“A machine learning and explainable ai approach for

predicting secondary school student performance,” in

2022 IEEE 12th Annual Computing and

Communication Workshop and Conference (CCWC),

2022, pp. 0399– 0405.

Kaihua Wei, Bojian Chen, Jingcheng Zhang, Shanhui Fan,

Kaihua Wu, Guangyu Liu and Dongmei Chen,

“Explainable Deep Learning Study for Leaf Disease

Classification”, Agronomy 2022, 12, 1035,

https://doi.org/ 10.3390/agronomy12051035

Ko Ko Zaw, Dr. Zin Ma Ma Myo, Daw Thae Hsu Thoung,

“Support Vector Machine Based Classification of Leaf

Diseases”, International Journal of Science and

Engineering Applications, 2018.

Quan Huu Cap et al. “LeafGAN: An Effective Data

Augmentation Method for Practical Plant Disease

Diagnosis”. In: IEEE Transactions on Automation

Science and Engineering 19.2 (Apr. 2022). Conference

Name: IEEE Transactions on Automation Science and

Engineering, pp. 1258–1267. ISSN: 1558-3783.

S. M. Lundberg and S. I. Lee, “A unified approach to

interpreting model predictions,” Adv. Neural Inf.

Process. Syst., vol. 2017-Decem, no. Section 2, pp.

4766–4775, 2017

Uday Pratap Singh et al. “Multilayer Convolution Neural

Network for the Classification of Mango Leaves

Infected by Anthracnose Disease”. In: IEEE Access 7

(2019). Conference Name: IEEE Access, pp. 43721–

43729.

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

744