深度感測技術在復建的使用情境User Senarios of Depth Sensing in Rehabilitation

比爾蓋茲說過,人們高估了新科技在兩年內的可能,但卻低估了他在十年後的應用性

深度感測器相關的應用,或者如何讓電腦"看"的懂人們的動作與姿態,進而做科學與醫學上的分析解構一直都是個研究與創業上相當有趣的挑戰。從重裝備的VICON-全身貼滿光球,精準的抓取使用者在空間中隨著時間流氅的每個精確的資訊。抑或,想要更有效與電腦互動,2005卡內基美隆的一位博士生John將wii的手把改裝作為更為直覺的互動工具。都初步看出了這樣技術對於改變人機互動的可能性。

一直到初代Kinect 360問世之後,才輕易的將這兩個問題有效的解決。能夠在單一感測器的前提下,完成了全身姿態的動態量測,但又能夠有效的與電腦或任何數位化設備互動。爾後的十幾年間,隨著軟體技術與硬體技術的成熟,包括了更精確的空間解析度(可以看得更準確),或是適用於更繁複的場景(AI訓練模型的優化),抑或運算的效率讓行動裝置也能夠簡易使用。

在這樣的基礎下,龍骨王將其應用在復健上遇到的問題,而進一步的解決甚至是創造了應用的可行性。包括了產品銷售至百餘家醫院/機構,每年使用產品來協助與回復健康近十萬人次的數量。我們綜整了目前的幾種應用情境

醫院/機構內

在專屬的應用空間,目前在醫學中心只要有占地3*3公尺左右的訓練教室或是空間,就可以使用龍骨王的產品進行復健。

也有些患者不習慣或是因為任何原因需要在自己病床獨自進行復健,使用龍骨王的行動版本的產品也都能進行。

居家復健

原本設想的是B2B2C讓醫院開立處方箋讓患者帶回使用。使用期程為兩周為一個療程。

同時間,各國也有Traveling Physiotherapist, 也就是到府的治療師/照服人員,由於他們兼具了醫療各方面的任務,帶著開立的課程到案家家中,也是目前常見的應用。

利用步態早期檢測失智症?Unveiling the Link Between Gait and Dementia: A Path Towards Early Detection

在醫學研究的領域裡,每一次突破都承載著改變生活的承諾與可能性。今天,我們將深入探討一項深刻的研究,揭示了步態異常與失智症發作之間錯綜複雜的關係。作為數位醫療領域的創業者,這一發現為步態分析提供了針對目標客戶沉痾已久的問題,有望展開革新失智症的早期檢測方法。

傳統上,對於失智症的預臨床評估主要圍繞在認知測量上,尤其是關注記憶功能,通常是為了預測阿茲海默症。然而,這來自澳洲老年人研究擴展了我們的視角,將步態分析這個非認知的臨床指標,納入了預臨床定義中,顯著提高了其對於失智症的預測能力。

這項開創性研究追蹤了630名75歲及以上的社區居民六年,將他們分為三組:失智、認知完好或可能表現出阿茲海默症、血管性失智症、錐體外失智症或以上各種組合的預臨床階段。值得注意的是,偵測到步態和運動減緩相結合的認知障礙參與者,被確定為最容易在研究期間(六年)內發展為失智症的人群(OR 5.6; 95% CI 2.5-12.6)

此外,研究結果挑戰了傳統關於MRI掃描白質指數與步態異常之間相關性的假設,揭示了不一致的聯繫。這凸顯了在臨床環境中需要更簡單、更易於獲取的步態評估工具的必要性。

事實上,這項研究的影響深遠。步態的簡單測量可能作為助益良多的臨床工具,為醫生提供了預測失智症發作的另一個維度。然而,必須承認,我們仍需努力去揭示這些缺陷的潛在本質。以龍骨王的角度,我們準備迎接這一挑戰,推動創新,解析步態異常的複雜性及其對失智症的影響。

總之,這項研究預示著下一個階段的機會。通過利用新穎的技術和跨學科合作,我們可以充分利用步態分析的潛力,重新定義失智症檢測和干預的格局。



In the realm of medical research, every breakthrough holds the promise of transforming lives. Today, we delve into a profound study shedding light on the intricate relationship between gait abnormalities and the onset of dementia. As an aspiring entrepreneur in the healthcare sector, this revelation paves the way for innovative solutions in gait analysis, with the potential to revolutionize early detection methods for dementia.

Traditionally, assessments for preclinical dementia predominantly revolved around cognitive measures, particularly focusing on memory functions, often to predict Alzheimer’s disease. However, the Sydney Older Persons Study has widened our perspective by incorporating non-cognitive clinical markers into preclinical definitions, enhancing their predictive power significantly.

This groundbreaking study followed 630 community-dwelling participants aged 75 or older for six years, categorizing them into groups: demented, cognitively intact, or possibly exhibiting preclinical phases of Alzheimer’s disease, vascular dementia, extrapyramidal dementia, or various combinations thereof. Remarkably, participants displaying cognitive impairment alongside gait and motor slowing were identified as the most susceptible to dementia progression over the study period, with odds ratios (OR) of 5.6 and 3.3 for dementia incidence and mortality, respectively.

Moreover, the study findings challenge conventional assumptions regarding the correlation between white matter indices on MRI scanning and gait abnormalities, revealing inconsistent associations. This underscores the need for simpler, more accessible tools for gait assessment in clinical settings.

Indeed, the implications of this research are profound. Simple measures of gait may serve as invaluable clinical tools, offering clinicians an additional dimension for predicting dementia onset. However, it’s crucial to acknowledge the ongoing quest to uncover the underlying nature of these deficits. As enthusiastic entrepreneurs in the healthcare landscape, we’re poised to embrace this challenge, driving innovation to decipher the complexities of gait abnormalities and their implications for dementia.

In essence, this study ignites a spark of inspiration within the entrepreneurial community, signaling a call to action. By leveraging cutting-edge technologies and interdisciplinary collaborations, we can harness the potential of gait analysis to redefine the landscape of dementia detection and intervention.

Verghese, J., Derby, C., Katz, M., & Lipton, R. (2007). High risk neurological gait syndrome and vascular dementia. Journal of Neural Transmission, 114, 1249-1252. https://doi.org/10.1007/s00702-007-0762-0.

神的恩典-步態分析

神之步態

這是一位家醫科醫師的口述,我們在海上衝浪時聊天的內容。

那是一間醫院的家醫科,不確定那時是醫學中心或是類醫學中心。有一次,應科內的要求,到醫院附近的教會參與活動,做好鄰里關係為後續的後送或是患者來源也是相當重要,這位浪友補充到。

教會所在的地方還是高齡者為主,教會為他們的教友準備了輕便的步態量測,每次參加聚會、禱告都可以在神的恩典下獲得免費的步態分析。只要簡單的幾秒就可以….

聽到這邊有一道分岔的大浪過來,我們就快速的分頭去追自己的浪頭。海上的談話總是這麼的用心傾聽卻又容易被浪頭吸引走。

這的確是龍骨王的產品步態分析,我依稀記得同仁有賣過宗教體系的客戶,但後續產品怎麼被使用,倒是沒有詳細的追蹤。原本在醫院系統內昂貴而使用繁複的產品,輕量化之後運用在各個機構/社區之後,各樣的使用倒是遠遠超出了一開始的預期。

就像是普羅米修斯為人們帶來的火種,改變了生活許多面貌,倒是沒有預期後面會有這麼多有趣的事情發生。

或許應該回過頭再去拜訪採購的客戶們,看看他們使用的狀況,在新的一年裡。

運用龍骨王的訓練產品來解決香港高齡化的問題


香港理工大學(理大)上週五(11 月 3 日)舉辦「賽馬會智齡匯」計劃成果分享會,我們香港的夥伴Kerry Medical受邀在論壇中分享HappyGoGo在「賽馬會智齡匯」示範院舍的應用與對長者們的幫助。
香港在政府、地方單位的支持下,許多長照機構非常幸福可以導入各式各樣的科技產品來提升照護品質,很多前瞻的長照驗證計畫也都在香港如火如荼地展開。

今年香港社會服務聯會(HKCSS)亦開始了為期兩年的居家復健實證計畫,連結多間的治療所、多位治療師,共同通過數位復健系統將復健服務推送到居家。一舉解決了高齡化的臨床需求而復健醫療人力無法全面覆蓋的問題,科技化恰恰是解決該方法的首要工具。香港政府為高齡化的社會展示了導入科技、善用科技來在對長者的健康關懷中做出最佳的示範。

“復健的目的,就是讓患者回到原本的生活"-陽明交通大學物理治療系教授楊雅如老師

HappyGoGo最初設計的目的,就是以原有復健訓練目的為核心,藉由超過兩百款各式各樣的生活情境讓使用者逐步地建立信心能夠有能力康復回到原本的生活情境,添加了遊戲的趣味性與競爭性,讓患者/長者能夠更加地投入增加使用的時間,進而更有效率的回復。而這樣的產品也經歷台灣各級醫院/機構乃至居家的使用,與相關論文的發表驗證其成效。


姿勢評価:指標と一般的な原因

体の姿勢の評価は、いくつかの側面から観察できます。

頭の傾き: 頭の傾きは、片側の首の筋肉が他の側よりも緊張しているためかもしれません。このような姿勢を長時間続けると、首のこりや痛み、頭痛の原因となる可能性があります。頻繁に携帯電話をチェックすることや、画面の位置が不正確であることが、一般的な原因として考えられます。

肩の水平: 肩が不均衡であることは、一方の肩や上背部の筋肉が緊張していることを示している可能性があります。この姿勢を長時間続けると、背中、肩、首の痛みの原因となる可能性があります。長時間のコンピュータの使用、重い物を持ち上げること、または不均衡な運動トレーニングが一般的な原因として考えられます。

腕の角度: 腕が過度に曲がっている場合、それは肘や手首の過度な使用を示している可能性があり、筋肉の痛みやこりの原因となる可能性があります。長時間のタイピング、筆記、または他の繰り返しの動きが原因となる可能性があります。

足首の距離: 足首の距離が狭すぎると、立っているときの安定性に影響を与え、転倒のリスクが増加する可能性があります。足首の距離が広すぎると、下肢の圧力分布が不均衡となり、脚の痛みの原因となる可能性があります。

骨盤の傾き: 骨盤の不均衡は、一方の臀部や腰の筋肉が緊張していることを示している可能性があります。これは、腰の痛みや下肢の機能障害の原因となる可能性があります。不正確な座り方、長時間片側に立っていること、または足の怪我が一般的な原因として考えられます。

胴体の傾きの角度: 胴体が前傾している場合、背部の筋肉が弱く、腹部の筋肉が緊張していることを示している可能性があります。胴体が後傾している場合、腹部の筋肉が弱く、背部の筋肉が緊張していることを示している可能性があります。これらの状況は、腰の痛みや不快感の原因となる可能性があります。長時間の座り方、不正確な持ち上げ技術、または筋肉の不均衡が一般的な原因として考えられます。

人の観察は結局のところ主観的であり、経験の積み重ねが必要です。しかし、コンピュータビジョンAIの支援を受けることで、詳細を正確に観察し、値を直接デジタル化することができます。

もちろん、上記の評価は、体の姿勢に関するいくつかの基本的な指示を提供します。前述

Posture Assessment: Indicators and Common Causes

Assessment of body posture can be observed from several aspects:

Head tilt: Head tilt might be due to one side of the neck muscles being tighter than the other, causing the head to lean to one side. Prolonged posture like this may cause neck stiffness, pain, or headaches. Frequently checking phones or incorrect screen positions may be common reasons.

Shoulder level: Unbalanced shoulders might indicate that the muscles of one side of the shoulder or upper back are tighter. Maintaining this posture for a long time can lead to back, shoulder, or neck pain. Prolonged computer use, lifting heavy objects, or unbalanced exercise training could be common reasons.

Arm angle: Excessively bent arms might indicate overuse of the elbows or wrists, which can lead to muscle pain or stiffness. Typing, writing, or other repetitive motions for long durations might be the cause.

Ankle distance: A too small distance between the ankles might affect standing stability, increasing the risk of falls. Too large an ankle gap might result in uneven pressure distribution in the lower limbs, possibly causing leg pain.

Pelvic tilt: An imbalance in the pelvis might indicate that the muscles on one side of the buttocks or waist are tighter. This could result in lower back pain or functional disorders of the lower limbs. Incorrect sitting posture, prolonged standing on one side, or foot injuries might be common causes.

Trunk tilt angle: A forward-leaning trunk might indicate weakened back muscles, with tighter abdominal muscles. A backward-leaning trunk might indicate weakened abdominal muscles and tighter back muscles. Both situations might cause lower back pain or discomfort. Prolonged sitting, incorrect lifting techniques, or muscle imbalances could be common reasons.

After all, human observation is subjective and requires the accumulation of experience. But with the assistance of computer vision AI, details can be observed accurately, and values can be directly digitized.

Of course, the above evaluations provide some basic indications of body posture. As mentioned, to determine specific health issues or symptoms, a more in-depth clinical assessment is needed. If there are symptoms of pain or discomfort, it is recommended to consult a doctor or therapist early.

AI 應用於復健/高齡

這是ai人工智慧學校與產業的的意見領袖appWorks等專家團隊遴選出台灣AI在各個垂直領域的廠商。龍骨王很榮幸的名列其中。這當中有個論述。

1.導入AI就開了外掛無敵了?

大約是2016年AI在矽谷大爆發,2017年那時進入UCBerkeley當訪問學者可真是開了眼界,不論是指導教授的專題研討時各方聰明的方式,或是研討會/新創公司都各有巧妙之處。但真的到AI成了新的顯學這的確是chatGPT出來之後的事情。在當時就開始將新型的AI影像偵測技術以及資料分類方式導入產品內。產品醫療的宣稱要能符合臨床需求,驗證型的論文要具科學說服力。更高的精確度/成本效益/更強力的臨床宣稱,但面向客戶的產品端要能極端的易懂而操作簡易,以及,你的客戶購買導入你的產品後怎麼產生更高的可見效益(賺更多錢)。這些是導入AI之後依樣要面對的商業挑戰,而非導入AI萬事俱矣。

2.2030已經沒有AI公司!

就像是當年電子商務即將崛起的千禧年前夕,Intel CEO曾經說過的預測,未來將沒有網路公司。是絕大部分的公司都會導入網路應用。AI也是核心提升的一個階段,其效益可以類比於工業革命之於農業社會。再加上科技迭代的速度增快,屆時的AI使用,可能跟現在你開瀏覽器上網或是使用word記錄公司會議紀錄或用excel快速拉一個成本效益表,對一間公司的運營來說一樣簡單或必要。

3.我該怎麼導入AI?

產品型/服務型的公司,或是原有就配置有研發工程師的公司,就從現有的服務思考可以怎麼樣利用AI來達成十倍或是百倍效益。或是,跟圖列表中的這些公司當好朋友:)

遠距復健的執行模式

上了湘雲海上衝浪私人教練課程,發現善用科技的特點,而這跟遠距復健有異曲同工之妙。在岸上敘述完基礎理論時,下海他用高倍率的鏡頭以及防水的耳機麥克風,指導我追浪下浪時所需要的肢體動作特點。

衝浪相對而言是多個複雜動作流暢組合起來運動,還要對應每次大小強度都不同的浪況。即使對於原理再怎麼理解通徹,在海上遇到不同的波型,大腦聽懂但身體很難自動地跟上,尤其在有限的反應時間內。這次有湘雲在岸上觀察並且用耳麥提點我那些浪可以注意,划水動作可以修正(即時的回饋),或是追到浪之後應該要馬上挺身重心轉移,由於當下還沒練成肌肉記憶,很多動作的組合都需要靠大腦處理,還好有教練即時的給予回饋與提醒。

回到岸上之後,我們再根據即時的錄影,回播每一個動作詳細解說。

—這完全是遠距復健的模式—-

患者回到家中之後需要持續型的復健目前的做法大抵有三種

1.請全職的物理治療師每天到府陪伴三小時

就像是我的教練自己會到海上跟著我的浪板一起,跟我一起衝,這模式比較像是我在印尼衝浪的模式。但這就考慮到教練的成本,以及效率。從遠端利用科技觀察或許能看出更全面的問題癥結。

2.治療師以設備觀測,患者兩周回醫院看診一次

教練在海上以鏡頭(自動)的觀測動作,選手上岸之後再根據記錄到的內容進行分析與講解。缺點是沒有辦法即時的更正動作,但相對來說成本最優

3.治療師以設備觀測,且在線上以語音即時給予回饋。

就是這次所提及的方式,效果好,但是成本來說還是偏高。

目前在台灣只能打高端客群或是歐美客戶。但不失是個好方法。

Patient who needs GaitAnalysis, and how to evaluate

Musculoskeletal conditions:
Patients with musculoskeletal conditions such as osteoarthritis or back pain may benefit from gait analysis to identify issues such as poor posture or gait imbalances that could be contributing to their condition. Gait analysis can also help determine whether a patient’s walking or running technique may be putting excess stress on joints or muscles, which could be contributing to pain.

Musculoskeletal conditions can affect a person’s gait in various ways, depending on the specific condition and its severity. However, some common gait parameters that may indicate abnormalities in individuals with musculoskeletal conditions are:

Step length: Individuals with musculoskeletal conditions may have reduced step length on the affected side due to pain or weakness.

Cadence: Cadence refers to the number of steps taken per minute. Individuals with musculoskeletal conditions may have a slower cadence due to pain or difficulty with movement.

Foot strike: Foot strike refers to the point at which the foot makes contact with the ground. Individuals with musculoskeletal conditions may have an altered foot strike pattern due to pain or limited range of motion.

Joint angles: Musculoskeletal conditions can affect joint range of motion, leading to altered joint angles during gait. For example, individuals with hip or knee osteoarthritis may have reduced hip or knee flexion during gait.

Ground reaction forces: Musculoskeletal conditions can alter the distribution and magnitude of forces acting on the foot and lower extremity during gait. Gait analysis can measure these forces to identify abnormalities and guide treatment.

Neurological conditions:
Patients with neurological conditions such as Parkinson’s disease or multiple sclerosis may experience gait abnormalities that can affect their mobility and increase the risk of falls. Gait analysis can help identify these abnormalities and develop treatment plans to improve gait and reduce the risk of falls. For example, gait analysis may help determine whether a patient would benefit from assistive devices such as a cane or walker, or whether specific exercises or physical therapy may be helpful.

some common gait parameters that may indicate abnormalities in individuals with neurological conditions are:

Step length: Individuals with neurological conditions may have a reduced step length due to weakness, spasticity, or ataxia.

Cadence: Individuals with neurological conditions may have a slower or variable cadence due to difficulty with movement control.

Foot drop: Foot drop refers to the inability to lift the foot during the swing phase of gait, which can result in tripping or falling. Individuals with neurological conditions may have foot drop due to weakness or spasticity.

Timing and coordination: Neurological conditions can affect the timing and coordination of gait, leading to an abnormal gait pattern. For example, individuals with Parkinson’s disease may have a shuffling gait and reduced arm swing due to bradykinesia and rigidity.

Gait variability: Individuals with neurological conditions may have increased gait variability, meaning that their gait parameters may vary more than usual from step to step. This can be indicative of instability or impaired motor control.

Injury recovery:
Patients recovering from an injury such as a knee or hip replacement may benefit from gait analysis to monitor their progress and ensure that they are moving correctly to avoid further injury. Gait analysis can also help identify any issues with mobility or gait that may be hindering recovery, so that appropriate interventions can be made.

Weight-bearing asymmetry: Injury recovery patients may have reduced weight-bearing on the affected limb or side, resulting in reduced step length and cadence on that side.

Kinematic changes: Injury recovery patients may have altered joint angles and ranges of motion during gait due to compensatory movements or changes in biomechanics.

Muscle activation patterns: Injury recovery patients may have altered muscle activation patterns during gait due to changes in neuromuscular control.

Ground reaction forces: Injury recovery patients may have altered ground reaction forces during gait due to changes in weight-bearing or altered biomechanics.

Pain: Pain can significantly affect gait parameters in injury recovery patients, leading to altered step length, cadence, and joint kinematics.

Children:
Children with gait abnormalities such as toe walking or flat feet may benefit from gait analysis to determine the cause of their condition and develop appropriate interventions. For example, gait analysis may help determine whether physical therapy, orthotics, or other interventions may be helpful in improving gait and reducing the risk of future issues.

gait analysis is typically used to identify abnormalities or deviations from normal development. Some common gait parameters that may indicate abnormalities in children are:

Step length: Children may have reduced step length due to neuromuscular or skeletal abnormalities.

Cadence: Children may have a slower or variable cadence due to developmental delays or neuromuscular disorders.

Foot position: Children may have an abnormal foot position during gait, such as toe-walking or in-toeing, which may be indicative of neuromuscular or skeletal abnormalities.

Joint angles: Children may have abnormal joint angles during gait due to neuromuscular or skeletal abnormalities.

Muscle activation patterns: Children may have abnormal muscle activation patterns during gait due to neuromuscular disorders or developmental delays.

Overall, gait analysis can be a valuable tool in identifying issues with gait and mobility in patients with various medical conditions, and developing appropriate treatment plans to improve function and quality of life.

It’s important to note that gait analysis is individualized and tailored to each patient’s specific condition and needs. A thorough evaluation by a qualified healthcare professional is necessary to determine which gait parameters are most relevant for each individual patient.

Overview of Gait Analysis Techniques and Their Applications

There are several different technical principles used in gait analysis, including:

  1. Motion capture systems:
    Using multiple cameras and sensors to capture three-dimensional position and posture data of human movement, analyzing gait features and abnormalities. moden technology enable a single depthsensor camera to capture necessary three-dimensional position and posture, it is convenient and easy to use.
  2. Pressure-sensitive floors:
    Placing pressure sensors on the floor to measure foot pressure distribution at different parts, analyzing gait features and abnormalities. Some more information may be provided such as shear force. This is widely used in research laboratory in Taiwan.
  3. Accelerometers and gyroscopes(wearable):
    Measuring body motion acceleration and angular velocity by placing sensors such as accelerometers and gyroscopes at different body parts, analyzing gait features and abnormalities. It takes some time and experience in installation. Some patient might be fatigue during the process, under the condition of have full body information.
  4. Electromyography (EMG): Measuring the electrical signals of muscle contraction and relaxation by placing electrodes on the muscle, analyzing gait features and abnormalities.
  5. Single RGB camera:
    Using a regular 2D camera, such as those found on mobile phones, and the latest computer vision AI algorithms to analyze posture. By assigning distances in the real world and deducing the possible distance from the camera via the ratio between selected joint distances, gait features and abnormalities can be analyzed. This method is low-cost and convenient, but may have limitations in accurately measuring gait parameters due to the lack of concrete 3D information.

Each method has its advantages and disadvantages. For example, motion capture systems are considered the most accurate method for measuring gait, but they are expensive and require specialized equipment and trained personnel. Pressure-sensitive floors are more affordable but have limited portability and may not be suitable for outdoor use. Accelerometers and gyroscopes are relatively low-cost and portable but have limitations in measuring gait abnormalities. EMG can provide information about muscle activity, but it requires placing electrodes on the skin, which can be uncomfortable for some people. Infrared sensors are non-intrusive and easy to use but have limitations in measuring certain gait parameters.

Overall, the choice of method depends on the specific requirements of the gait analysis application and the trade-offs between accuracy, cost, portability, and ease of use.

In conclusion, gait analysis is an important tool in assessing and managing gait abnormalities and related conditions. There are various technical principles used in gait analysis, each with its own advantages and limitations. The choice of method depends on the specific needs of the application and the trade-offs between accuracy, cost, portability, and ease of use.

At LongGood MediTech, we are committed to providing cutting-edge digital rehabilitation solutions for our customers. Our flagship product, GaitBEST, is a motion capture system designed specifically for gait analysis. It offers accurate and reliable results, while remaining affordable and user-friendly, making it an ideal solution for hospitals, elder care centers, and other healthcare facilities.

Our team of experts is dedicated to providing excellent customer support and ensuring that our products meet the needs of our clients. We believe that by providing accessible and effective digital rehabilitation solutions, we can make a positive impact on the lives of individuals with gait abnormalities and related conditions.

With our expertise and commitment to innovation, LongGood MediTech is poised to lead the way in digital rehabilitation and gait analysis. Contact us today to learn more about our products and services.